Merge pull request #13887 from open-webui/dev

0.6.10
Update CHANGELOG.md
2025-05-19 05:34:23 +04:00 · 2025-05-19 05:27:29 +04:00 · 2025-05-19 05:24:03 +04:00 · 2025-05-19 05:20:57 +04:00 · 2025-05-19 05:12:53 +04:00 · 2025-05-19 04:33:34 +04:00
281 changed files with 21994 additions and 8826 deletions
@@ -70,3 +70,7 @@
 ### Screenshots or Videos

 - [Attach any relevant screenshots or videos demonstrating the changes]
+
+### Contributor License Agreement
+
+By submitting this pull request, I confirm that I have read and fully agree to the [Contributor License Agreement (CLA)](/CONTRIBUTOR_LICENSE_AGREEMENT), and I am providing my contributions under its terms.
@@ -5,6 +5,165 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

+## [0.6.10] - 2025-05-19
+
+### Added
+
+- 🧩 **Experimental Azure OpenAI Support**: Instantly connect to Azure OpenAI endpoints by simply pasting your Azure OpenAI URL into the model connections—bringing flexible, enterprise-grade AI integration directly to your workflow.
+- 💧 **Watermark AI Responses**: Easily add a visible watermark to AI-generated responses for clear content provenance and compliance with EU AI regulations—perfect for regulated environments and transparent communication.
+- 🔍 **Enhanced Search Experience with Dedicated Modal**: Enjoy a modern, powerful search UI in a dedicated modal (open with Ctrl/Cmd + K) accessible from anywhere—quickly find chats, models, or content and boost your productivity.
+- 🔲 **"Toggle" Filter Type for Chat Input**: Add interactive toggle filters (e.g. Web Search, Image, Code Interpreter) right into the chat input—giving you one-click control to activate features and simplifying the chat configuration process.
+- 🧰 **Granular Model Capabilities Editor**: Define detailed capabilities and feature access for each AI model directly from the model editor—enabling tailored behavior, modular control, and a more customized AI environment for every team or use case.
+- 🌐 **Flexible Local and External Connection Support**: Now you can designate any AI connection—whether OpenAI, Ollama, or others—as local or external, enabling seamless setup for on-premise, self-hosted, or cloud configurations and giving you maximum control and flexibility.
+- 🗂️ **Allowed File Extensions for RAG**: Gain full control over your Retrieval-Augmented Generation (RAG) workflows by specifying exactly which file extensions are permitted for upload, improving security and relevance of indexed documents.
+- 🔊 **Enhanced Audio Transcription Logic**: Experience smoother, more reliable audio transcription—very long audio files are now automatically split and processed in segments, preventing errors and ensuring even challenging files are transcribed seamlessly, all part of a broader stability upgrade for robust media workflows.
+- 🦾 **External Document Loader Support**: Enhance knowledge base building by integrating documents using external loaders from a wide range of data sources, expanding what your AI can read and process.
+- 📝 **Preview Button for Code Artifacts**: Instantly jump from an HTML code block to its associated artifacts page with the click of a new preview button—speeding up review and streamlining analysis.
+- 🦻 **Screen Reader Support for Response Messages**: All chat responses are now fully compatible with screen readers, making the platform more inclusive and accessible for everyone.
+- 🧑‍💼 **Customizable Pending User Overlay**: You can now tailor the overlay title and content shown to pending users, ensuring onboarding messaging is perfectly aligned with your organization’s tone and requirements.
+- 🔐 **Option to Disable LDAP Certificate Validation**: You can now disable LDAP certificate validation for maximum flexibility in diverse IT environments—making integrations and troubleshooting much easier.
+- 🎯 **Workspace Search by Username or Email**: Easily search across workspace pages using any username or email address, streamlining user and resource management.
+- 🎨 **High Contrast & Dark Mode Enhancements**: Further improved placeholder, input, suggestion, toast, and model selector contrasts—including a dedicated placeholder dark mode—for more comfortable viewing in all lighting conditions.
+- 🛡️ **Refined Security for Pipelines & Model Uploads**: Strengthened safeguards against path traversal vulnerabilities during uploads—ensuring your platform’s document and model management remains secure.
+- 🌏 **Major Localization Upgrades**: Comprehensive translation updates and improvements across Korean, Bulgarian, Catalan, Japanese, Italian, Traditional Chinese, and Spanish—including more accurate AI terminology for clarity; your experience is now more natural, inclusive, and professional everywhere.
+- 🦾 **General Backend Stability & Security**: Multiple backend improvements (including file upload, command navigation, and logging refactorings) deliver increased resilience, better error handling, and a more robust platform for all users.
+
+### Fixed
+
+- ✅ **Evaluation Feedback Endpoint Reliability**: Addressed issues with feedback submission endpoints to ensure seamless user feedback collection on model responses.
+- 🫰 **Model List State Fixes**: Resolved issues where model status toggles in the workspace page might inadvertently switch or confuse state, making the management of active/inactive models more dependable.
+- ✍️ **Admin Signup Logic**: Admin self-signup experience and validation flow is smoother and more robust.
+- 🔁 **Signout Redirect Flow Improved**: Logging out now redirects more reliably, reducing confusion and making session management seamless.
+
+## [0.6.9] - 2025-05-10
+
+### Added
+
+- 📝 **Edit Attached Images/Files in Messages**: You can now easily edit your sent messages by removing attached files—streamlining document management, correcting mistakes on the fly, and keeping your chats clutter-free.
+- 🚨 **Clear Alerts for Private Task Models**: When interacting with private task models, the UI now clearly alerts you—making it easier to understand resource availability and access, reducing confusion during workflow setup.
+
+### Fixed
+
+- 🛡️ **Confirm Dialog Focus Trap Reliability**: The focus now stays correctly within confirmation dialogs, ensuring keyboard navigation and accessibility is seamless and preventing accidental operations—especially helpful during critical or rapid workflows.
+- 💬 **Temporary Chat Admin Controls & Session Cleanliness**: Admins are now able to properly enable temporary chat mode without errors, and previous session prompts or tool selections no longer carry over—delivering a fresh, predictable, and consistent temporary chat experience every time.
+- 🤖 **External Reranker Integration Functionality Restored**: External reranker integrations now work correctly, allowing you to fully leverage advanced ranking services for sharper, more relevant search results in your RAG and knowledge base workflows.
+
+## [0.6.8] - 2025-05-10
+
+### Added
+
+- 🏆 **External Reranker Support for Knowledge Base Search**: Supercharge your Retrieval-Augmented Generation (RAG) workflows with the new External Reranker integration; easily plug in advanced reranking services via the UI to deliver sharper and more relevant search results, accelerating research and insight discovery.
+- 📤 **Unstylized PDF Export Option (Reduced File Size)**: When exporting chat transcripts or documents, you can now choose an unstylized PDF export for snappier downloads, minimal file size, and clean data archiving—perfect for large-scale storage or sharing.
+- 📝 **Vazirmatn Font for Persian & Arabic**: Arabic and Persian users will now see their text beautifully rendered with the specialized Vazirmatn font for an improved localized reading experience.
+- 🏷️ **SharePoint Tenant ID Support for OneDrive**: You can now specify a SharePoint tenant ID in OneDrive settings for seamless authentication and granular enterprise integration.
+- 👤 **Refresh OAuth Profile Picture**: Your OAuth profile picture now updates in real-time, ensuring your presence and avatar always match your latest identity across integrated platforms.
+- 🔧 **Milvus Configuration Improvements**: Configure index and metric types for Milvus directly within settings; take full control of your vector database for more accurate and robust AI search experiences.
+- 🛡️ **S3 Tagging Toggle for Compatibility**: Optional S3 tagging via an environment toggle grants full compatibility with all storage backends—including those that don’t support tagging like Cloudflare R2—ensuring error-free attachment and document management.
+- 👨‍🦯 **Icon Button Accessibility Improvements**: Key interactive icon-buttons now include aria-labels and ARIA descriptions, so screen readers provide precise guidance about what action each button performs for improved accessibility.
+- ♿ **Enhanced Accessibility with Modal Focus Trap**: Modal dialogs and pop-ups now feature a focus trap and improved ARIA roles, ensuring seamless navigation and screen reader support—making the interface friendlier for everyone, including keyboard and assistive tech users.
+- 🏃 **Improved Admin User List Loading Indicator**: The user list loading experience is now clearer and more responsive in the admin panel.
+- 🧑‍🤝‍🧑 **Larger Admin User List Page Size**: Admins can now manage up to 30 users per page in the admin interface, drastically reducing pagination and making large user teams easier and faster to manage.
+- 🌠 **Default Code Interpreter Prompt Clarified**: The built-in code interpreter prompt is now more explicit, preventing AI from wrapping code in Markdown blocks when not needed—ensuring properly formatted code runs as intended every time.
+- 🧾 **Improved Default Title Generation Prompt Template**: Title generation now uses a robust template for reliable JSON output, improving chat organization and searchability.
+- 🔗 **Support Jupyter Notebooks with Non-Root Base URLs**: Notebook-based code execution now supports non-root deployed Jupyter servers, granting full flexibility for hybrid or multi-user setups.
+- 📰 **UI Scrollbar Always Visible for Overflow Tools**: When available tools overflow the display, the scrollbar is now always visible and there’s a handy "show all" toggle, making navigation of large toolsets snappier and more intuitive.
+- 🛠️ **General Backend Refactoring for Stability**: Multiple under-the-hood improvements have been made across backend components, ensuring smoother performance, fewer errors, and a more reliable overall experience for all users.
+- 🚀 **Optimized Web Search for Faster Results**: Web search speed and performance have been significantly enhanced, delivering answers and sources in record time to accelerate your research-heavy workflows.
+- 💡 **More Supported Languages**: Expanded language support ensures an even wider range of users can enjoy an intuitive and natural interface in their native tongue.
+
+### Fixed
+
+- 🏃‍♂️ **Exhausting Workers in Nginx Reverse Proxy Due to Websocket Fix**: Websocket sessions are now fully compatible behind Nginx, eliminating worker exhaustion and restoring 24/7 reliability for real-time chats even in complex deployments.
+- 🎤 **Audio Transcription Issue with OpenAI Resolved**: OpenAI-based audio transcription now handles WebM and newer formats without error, ensuring seamless voice-to-text workflows every time.
+- 👉 **Message Input RTL Issue Fixed**: The chat message input now displays correctly for right-to-left languages, creating a flawless typing and reading experience for Arabic, Hebrew, and more.
+- 🀄 **Katex: Proper Rendering of Chinese Characters Next to Math**: Math formulas now render perfectly even when directly adjacent to Chinese (CJK) characters, improving visual clarity for multilingual teams and cross-language documents.
+- 🔂 **Duplicate Web Search URLs Eliminated**: Search results now reliably filter out URL duplicates, so your knowledge and search citations are always clean, trimmed, and easy to review.
+- 📄 **Markdown Rendering Fixed in Knowledge Bases**: Markdown is now displayed correctly within knowledge bases, enabling better formatting and clarity of information-rich files.
+- 🗂️ **LDAP Import/Loading Issue Resolved**: LDAP user imports process correctly, ensuring smooth onboarding and access without interruption.
+- 🌎 **Pinecone Batch Operations and Async Safety**: All Pinecone operations (batch insert, upsert, delete) now run efficiently and safely in an async environment, boosting performance and preventing slowdowns in large-scale RAG jobs.
+
+## [0.6.7] - 2025-05-07
+
+### Added
+
+- 🌐 **Custom Azure TTS API URL Support Added**: You can now define a custom Azure Text-to-Speech endpoint—enabling flexibility for enterprise deployments and regional compliance.
+- ⚙️ **TOOL_SERVER_CONNECTIONS Environment Variable Suppor**: Easily configure and deploy tool servers via environment variables, streamlining setup and enabling faster enterprise provisioning.
+- 👥 **Enhanced OAuth Group Handling as String or List**: OAuth group data can now be passed as either a list or a comma-separated string, improving compatibility with varied identity provider formats and reducing onboarding friction.
+
+### Fixed
+
+- 🧠 **Embedding with Ollama Proxy Endpoints Restored**: Fixed an issue where missing API config broke embedding for proxied Ollama models—ensuring consistent performance and compatibility.
+- 🔐 **OIDC OAuth Login Issue Resolved**: Users can once again sign in seamlessly using OpenID Connect-based OAuth, eliminating login interruptions and improving reliability.
+- 📝 **Notes Feature Access Fixed for Non-Admins**: Fixed an issue preventing non-admin users from accessing the Notes feature, restoring full cross-role collaboration capabilities.
+- 🖼️ **Tika Loader Image Extraction Problem Resolved**: Ensured TikaLoader now processes 'extract_images' parameter correctly, restoring complete file extraction functionality in document workflows.
+- 🎨 **Automatic1111 Image Model Setting Applied Properly**: Fixed an issue where switching to a specific image model via the UI wasn’t reflected in generation, re-enabling full visual creativity control.
+- 🏷️ **Multiple XML Tags in Messages Now Parsed Correctly**: Fixed parsing issues when messages included multiple XML-style tags, ensuring clean and unbroken rendering of rich content in chats.
+- 🖌️ **OpenAI Image Generation Issues Resolved**: Resolved broken image output when using OpenAI’s image generation, ensuring fully functional visual creation workflows.
+- 🔎 **Tool Server Settings UI Privacy Restored**: Prevented restricted users from accessing tool server settings via search—restoring tight permissions control and safeguarding sensitive configurations.
+- 🎧 **WebM Audio Transcription Now Supported**: Fixed an issue where WebM files failed during audio transcription—these formats are now fully supported, ensuring smoother voice note workflows and broader file compatibility.
+
+## [0.6.6] - 2025-05-05
+
+### Added
+
+- 📝 **AI-Enhanced Notes (With Audio Transcription)**: Effortlessly create notes, attach meeting or voice audio, and let the AI instantly enhance, summarize, or refine your notes using audio transcriptions—making your documentation smarter, cleaner, and more insightful with minimal effort.
+- 🔊 **Meeting Audio Recording & Import**: Seamlessly record audio from your meetings or capture screen audio and attach it to your notes—making it easier to revisit, annotate, and extract insights from important discussions.
+- 📁 **Import Markdown Notes Effortlessly**: Bring your existing knowledge library into Open WebUI by importing your Markdown notes, so you can leverage all advanced note management and AI features right away.
+- 👥 **Notes Permissions by User Group**: Fine-tune access and editing rights for notes based on user roles or groups, so you can delegate writing or restrict sensitive information as needed.
+- ☁️ **OneDrive & SharePoint Integration**: Keep your content in sync by connecting notes and files directly with OneDrive or SharePoint—unlocking fast enterprise import/export and seamless collaboration with your existing workflows.
+- 🗂️ **Paginated User List in Admin Panel**: Effortlessly manage and search through large teams via the new paginated user list—saving time and streamlining user administration in big organizations.
+- 🕹️ **Granular Chat Share & Export Permissions**: Enjoy enhanced control over who can share or export chats, enabling tighter governance and privacy in team and enterprise settings.
+- 🛑 **User Role Change Confirmation Dialog**: Reduce accidental privilege changes with a required confirmation step before updating user roles—improving security and preventing costly mistakes in team management.
+- 🚨 **Audit Log for Failed Login Attempts**: Quickly detect unauthorized access attempts or troubleshoot user login problems with detailed logs of failed authentication right in the audit trail.
+- 💡 **Dedicated 'Generate Title' Button for Chats**: Swiftly organize every conversation—tap the new button to let AI create relevant, clear titles for all your chats, saving time and reducing clutter.
+- 💬 **Notification Sound Always-On Option**: Take control of your notifications by setting sound alerts to always play—helping you stay on top of important updates in busy environments.
+- 🆔 **S3 File Tagging Support**: Uploaded files to S3 now include tags for better organization, searching, and integration with your file management policies.
+- 🛡️ **OAuth Blocked Groups Support**: Gain more control over group-based access by explicitly blocking specified OAuth groups—ideal for complex identity or security requirements.
+- 🚀 **Optimized Faster Web Search & Multi-Threaded Queries**: Enjoy dramatically faster web search and RAG (retrieval augmented generation) with revamped multi-threaded search—get richer, more accurate results in less time.
+- 🔍 **All-Knowledge Parallel Search**: Searches across your entire knowledge base now happen in parallel even in non-hybrid mode, speeding up responses and improving knowledge accuracy for every question.
+- 🌐 **New Firecrawl & Yacy Web Search Integrations**: Expand your world of information with two new advanced search engines—Firecrawl for deeper web insight and Yacy for decentralized, privacy-friendly search capabilities.
+- 🧠 **Configurable Docling OCR Engine & Language**: Use environment variables to fine-tune Docling OCR engine and supported languages for smarter, more tailored document extraction and RAG workflows.
+- 🗝️ **Enhanced Sentence Transformers Configuration**: Added new environment variables for easier set up and advanced customization of Sentence Transformers—ensuring best fit for your embedding needs.
+- 🌲 **Pinecone Vector Database Integration**: Index, search, and manage knowledge at enterprise scale with full native support for Pinecone as your vector database—effortlessly handle even the biggest document sets.
+- 🔄 **Automatic Requirements Installation for Tools & Functions**: Never worry about lost dependencies on restart—external function and tool requirements are now auto-installed at boot, ensuring tools always “just work.”
+- 🔒 **Automatic Sign-Out on Token Expiry**: Security is smarter—users are now automatically logged out if their authentication token expires, protecting sensitive content and ensuring compliance without disruption.
+- 🎬 **Automatic YouTube Embed Detection**: Paste YouTube links and see instant in-chat video embeds—no more manual embedding, making knowledge sharing and media consumption even easier for every team.
+- 🔄 **Expanded Language & Locale Support**: Translations for Danish, French, Russian, Traditional Chinese, Simplified Chinese, Thai, Catalan, German, and Korean have been upgraded, offering smoother, more natural user experiences across the platform.
+
+### Fixed
+
+- 🔒 **Tighter HTML Token Security**: HTML rendering is now restricted to admin-uploaded tokens only, reducing any risk of XSS and keeping your data safe.
+- 🔐 **Refined HTML Security and Token Handling**: Further hardened how HTML tokens and content are handled, guaranteeing even stronger resistance to security vulnerabilities and attacks.
+- 🔏 **Correct Model Usage with Ollama Proxy Prefixes**: Enhanced model reference handling so proxied models in Ollama always download and run correctly—even when using custom prefixes.
+- 📥 **Video File Upload Handling**: Prevented video files from being misclassified as text, fixing bugs with uploads and ensuring media files work as expected.
+- 🔄 **No More Dependent WebSocket Sequential Delays**: Streamlined WebSocket operation to prevent delays and maintain snappy real-time collaboration, especially in multi-user environments.
+- 🛠️ **More Robust Action Module Execution**: Multiple actions in a module now trigger as designed, increasing automation and scripting flexibility.
+- 📧 **Notification Webhooks**: Ensured that notification webhooks are always sent for user events, even when the user isn’t currently active.
+- 🗂️ **Smarter Knowledge Base Reindexing**: Knowledge reindexing continues even when corrupt or missing collections are encountered, keeping your search features running reliably.
+- 🏷️ **User Import with Profile Images**: When importing users, their profile images now come along—making onboarding and collaboration visually clearer from day one.
+- 💬 **OpenAI o-Series Universal Support**: All OpenAI o-series models are now seamlessly recognized and supported, unlocking more advanced capabilities and model choices for every workflow.
+
+### Changed
+
+- 📜 **Custom License Update & Contributor Agreement**: Open WebUI now operates under a custom license with Contributor License Agreement required by default—see https://docs.openwebui.com/license/ for details, ensuring sustainable open innovation for the community.
+- 🔨 **CUDA Docker Images Updated to 12.8**: Upgraded CUDA image support for faster, more compatible model inference and futureproof GPU performance in your AI infrastructure.
+- 🧱 **General Backend Refactoring for Reliability**: Continuous stability improvements streamline backend logic, reduce errors, and lay a stronger foundation for the next wave of feature releases—all under the hood for a more dependable WebUI.
+
+## [0.6.5] - 2025-04-14
+
+### Added
+
+- 🛂 **Granular Voice Feature Permissions Per User Group**: Admins can now separately manage access to Speech-to-Text (record voice), Text-to-Speech (read aloud), and Tool Calls for each user group—giving teams tighter control over voice features and enhanced governance across roles.
+- 🗣️ **Toggle Voice Activity Detection (VAD) for Whisper STT**: New environment variable lets you enable/disable VAD filtering with built-in Whisper speech-to-text, giving you flexibility to optimize for different audio quality and response accuracy levels.
+- 📋 **Copy Formatted Response Mode**: You can now enable “Copy Formatted” in Settings > Interface to copy AI responses exactly as styled (with rich formatting, links, and structure preserved), making it faster and cleaner to paste into documents, emails, or reports.
+- ⚙️ **Backend Stability and Performance Enhancements**: General backend refactoring improves system resilience, consistency, and overall reliability—offering smoother performance across workflows whether chatting, generating media, or using external tools.
+- 🌎 **Translation Refinements Across Multiple Languages**: Updated translations deliver smoother language localization, clearer labels, and improved international usability throughout the UI—ensuring a better experience for non-English speakers.
+
+### Fixed
+
+- 🛠️ **LDAP Login Reliability Restored**: Resolved a critical issue where some LDAP setups failed due to attribute parsing—ensuring consistent, secure, and seamless user authentication across enterprise deployments.
+- 🖼️ **Image Generation in Temporary Chats Now Works Properly**: Fixed a bug where image outputs weren’t generated during temporary chats—visual content can now be used reliably in all chat modes without interruptions.
+
 ## [0.6.4] - 2025-04-12

 ### Fixed
@@ -2,13 +2,13 @@

 ## Our Pledge

-As members, contributors, and leaders of this community, we pledge to make participation in our open-source project a harassment-free experience for everyone, regardless of age, body size, visible or invisible disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socioeconomic status, nationality, personal appearance, race, religion, or sexual identity and orientation.
+As members, contributors, and leaders of this community, we pledge to make participation in our project a harassment-free experience for everyone, regardless of age, body size, visible or invisible disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socioeconomic status, nationality, personal appearance, race, religion, or sexual identity and orientation.

 We are committed to creating and maintaining an open, respectful, and professional environment where positive contributions and meaningful discussions can flourish. By participating in this project, you agree to uphold these values and align your behavior to the standards outlined in this Code of Conduct.

 ## Why These Standards Are Important

-Open-source projects rely on a community of volunteers dedicating their time, expertise, and effort toward a shared goal. These projects are inherently collaborative but also fragile, as the success of the project depends on the goodwill, energy, and productivity of those involved.
+Projects rely on a community of volunteers dedicating their time, expertise, and effort toward a shared goal. These projects are inherently collaborative but also fragile, as the success of the project depends on the goodwill, energy, and productivity of those involved.

 Maintaining a positive and respectful environment is essential to safeguarding the integrity of this project and protecting contributors' efforts. Behavior that disrupts this atmosphere—whether through hostility, entitlement, or unprofessional conduct—can severely harm the morale and productivity of the community. **Strict enforcement of these standards ensures a safe and supportive space for meaningful collaboration.**

@@ -79,7 +79,7 @@ This approach ensures that disruptive behaviors are addressed swiftly and decisi

 ## Why Zero Tolerance Is Necessary

-Open-source projects thrive on collaboration, goodwill, and mutual respect. Toxic behaviors—such as entitlement, hostility, or persistent negativity—threaten not just individual contributors but the health of the project as a whole. Allowing such behaviors to persist robs contributors of their time, energy, and enthusiasm for the work they do.
+Projects thrive on collaboration, goodwill, and mutual respect. Toxic behaviors—such as entitlement, hostility, or persistent negativity—threaten not just individual contributors but the health of the project as a whole. Allowing such behaviors to persist robs contributors of their time, energy, and enthusiasm for the work they do.

 By enforcing a zero-tolerance policy, we ensure that the community remains a safe, welcoming space for all participants. These measures are not about harshness—they are about protecting contributors and fostering a productive environment where innovation can thrive.

@@ -0,0 +1,7 @@
+# Open WebUI Contributor License Agreement
+
+By submitting my contributions to Open WebUI, I grant Open WebUI full freedom to use my work in any way they choose, under any terms they like, both now and in the future. This approach helps ensure the project remains unified, flexible, and easy to maintain, while empowering Open WebUI to respond quickly to the needs of its users and the wider community.
+
+Taking part in this process means my work can be seamlessly integrated and combined with others, ensuring longevity and adaptability for everyone who benefits from the Open WebUI project. This collaborative approach strengthens the project’s future and helps guarantee that improvements can always be shared and distributed in the most effective way possible.
+
+**_To the fullest extent permitted by law, my contributions are provided on an “as is” basis, with no warranties or guarantees of any kind, and I disclaim any liability for any issues or damages arising from their use or incorporation into the project, regardless of the type of legal claim._**
@@ -1,64 +0,0 @@
-# Run with
-#    caddy run --envfile ./example.env --config ./Caddyfile.localhost
-#
-# This is configured for
-#    - Automatic HTTPS (even for localhost)
-#    - Reverse Proxying to Ollama API Base URL (http://localhost:11434/api)
-#    - CORS
-#    - HTTP Basic Auth API Tokens (uncomment basicauth section)
-
-
-# CORS Preflight (OPTIONS) + Request (GET, POST, PATCH, PUT, DELETE)
-(cors-api) {
-	@match-cors-api-preflight method OPTIONS
-	handle @match-cors-api-preflight {
-		header {
-			Access-Control-Allow-Origin "{http.request.header.origin}"
-			Access-Control-Allow-Methods "GET, POST, PUT, PATCH, DELETE, OPTIONS"
-			Access-Control-Allow-Headers "Origin, Accept, Authorization, Content-Type, X-Requested-With"
-			Access-Control-Allow-Credentials "true"
-			Access-Control-Max-Age "3600"
-			defer
-		}
-		respond "" 204
-	}
-
-	@match-cors-api-request {
-		not {
-			header Origin "{http.request.scheme}://{http.request.host}"
-		}
-		header Origin "{http.request.header.origin}"
-	}
-	handle @match-cors-api-request {
-		header {
-			Access-Control-Allow-Origin "{http.request.header.origin}"
-			Access-Control-Allow-Methods "GET, POST, PUT, PATCH, DELETE, OPTIONS"
-			Access-Control-Allow-Headers "Origin, Accept, Authorization, Content-Type, X-Requested-With"
-			Access-Control-Allow-Credentials "true"
-			Access-Control-Max-Age "3600"
-			defer
-		}
-	}
-}
-
-# replace localhost with example.com or whatever
-localhost {
-	## HTTP Basic Auth
-	## (uncomment to enable)
-	# basicauth {
-	# 	# see .example.env for how to generate tokens
-	# 	{env.OLLAMA_API_ID} {env.OLLAMA_API_TOKEN_DIGEST}
-	# }
-
-	handle /api/* {
-		# Comment to disable CORS
-		import cors-api
-
-		reverse_proxy localhost:11434
-	}
-
-	# Same-Origin Static Web Server
-	file_server {
-		root ./build/
-	}
-}
@@ -4,7 +4,7 @@
 ARG USE_CUDA=false
 ARG USE_OLLAMA=false
 # Tested with cu117 for CUDA 11 and cu121 for CUDA 12 (default)
-ARG USE_CUDA_VER=cu121
+ARG USE_CUDA_VER=cu128
 # any sentence transformer model; models to use can be found at https://huggingface.co/models?library=sentence-transformers
 # Leaderboard: https://huggingface.co/spaces/mteb/leaderboard 
 # for better performance and multilangauge support use "intfloat/multilingual-e5-large" (~2.5GB) or "intfloat/multilingual-e5-base" (~1.5GB)
@@ -26,6 +26,9 @@ ARG BUILD_HASH

 WORKDIR /app

+# to store git revision in build
+RUN apk add --no-cache git
+
 COPY package.json package-lock.json ./
 RUN npm ci

@@ -1,4 +1,4 @@
-Copyright (c) 2023-2025 Timothy Jaeryang Baek
+Copyright (c) 2023-2025 Timothy Jaeryang Baek (Open WebUI)
 All rights reserved.

 Redistribution and use in source and binary forms, with or without
@@ -15,6 +15,12 @@ modification, are permitted provided that the following conditions are met:
   contributors may be used to endorse or promote products derived from
   this software without specific prior written permission.

+4. Notwithstanding any other provision of this License, and as a material condition of the rights granted herein, licensees are strictly prohibited from altering, removing, obscuring, or replacing any "Open WebUI" branding, including but not limited to the name, logo, or any visual, textual, or symbolic identifiers that distinguish the software and its interfaces, in any deployment or distribution, regardless of the number of users, except as explicitly set forth in Clauses 5 and 6 below.
+
+5. The branding restriction enumerated in Clause 4 shall not apply in the following limited circumstances: (i) deployments or distributions where the total number of end users (defined as individual natural persons with direct access to the application) does not exceed fifty (50) within any rolling thirty (30) day period; (ii) cases in which the licensee is an official contributor to the codebase—with a substantive code change successfully merged into the main branch of the official codebase maintained by the copyright holder—who has obtained specific prior written permission for branding adjustment from the copyright holder; or (iii) where the licensee has obtained a duly executed enterprise license expressly permitting such modification. For all other cases, any removal or alteration of the "Open WebUI" branding shall constitute a material breach of license.
+
+6. All code, modifications, or derivative works incorporated into this project prior to the incorporation of this branding clause remain licensed under the BSD 3-Clause License, and prior contributors retain all BSD-3 rights therein; if any such contributor requests the removal of their BSD-3-licensed code, the copyright holder will do so, and any replacement code will be licensed under the project's primary license then in effect. By contributing after this clause's adoption, you agree to the project's Contributor License Agreement (CLA) and to these updated terms for all new contributions.
+
 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
@@ -7,7 +7,6 @@
 ![GitHub language count](https://img.shields.io/github/languages/count/open-webui/open-webui)
 ![GitHub top language](https://img.shields.io/github/languages/top/open-webui/open-webui)
 ![GitHub last commit](https://img.shields.io/github/last-commit/open-webui/open-webui?color=red)
-![Hits](https://hits.seeyoufarm.com/api/count/incr/badge.svg?url=https%3A%2F%2Fgithub.com%2Follama-webui%2Follama-wbui&count_bg=%2379C83D&title_bg=%23555555&icon=&icon_color=%23E7E7E7&title=hits&edge_flat=false)
 [![Discord](https://img.shields.io/badge/Discord-Open_WebUI-blue?logo=discord&logoColor=white)](https://discord.gg/5rJgQTnV4s)
 [![](https://img.shields.io/static/v1?label=Sponsor&message=%E2%9D%A4&logo=GitHub&color=%23fe8e86)](https://github.com/sponsors/tjbck)

@@ -62,9 +61,26 @@ For more information, be sure to check out our [Open WebUI Documentation](https:

 Want to learn more about Open WebUI's features? Check out our [Open WebUI documentation](https://docs.openwebui.com/features) for a comprehensive overview!

-## 🔗 Also Check Out Open WebUI Community!
+## Sponsors 🙌

-Don't forget to explore our sibling project, [Open WebUI Community](https://openwebui.com/), where you can discover, download, and explore customized Modelfiles. Open WebUI Community offers a wide range of exciting possibilities for enhancing your chat interactions with Open WebUI! 🚀
+#### Emerald
+
+<table>
+  <tr>
+    <td>
+      <a href="https://n8n.io/" target="_blank">
+        <img src="https://docs.openwebui.com/sponsors/logos/n8n.png" alt="n8n" style="width: 8rem; height: 8rem; border-radius: .75rem;" />
+      </a>
+    </td>
+    <td>
+      Does your interface have a backend yet?<br>Try <a href="https://n8n.io/">n8n</a>
+    </td>
+  </tr>
+</table>
+
+---
+
+We are incredibly grateful for the generous support of our sponsors. Their contributions help us to maintain and improve our project, ensuring we can continue to deliver quality work to our community. Thank you!

 ## How to Install 🚀

@@ -206,7 +222,7 @@ Discover upcoming features on our roadmap in the [Open WebUI Documentation](http

 ## License 📜

-This project is licensed under the [BSD-3-Clause License](LICENSE) - see the [LICENSE](LICENSE) file for details. 📄
+This project is licensed under the [Open WebUI License](LICENSE), a revised BSD-3-Clause license. You receive all the same rights as the classic BSD-3 license: you can use, modify, and distribute the software, including in proprietary and commercial products, with minimal restrictions. The only additional requirement is to preserve the "Open WebUI" branding, as detailed in the LICENSE file. For full terms, see the [LICENSE](LICENSE) document. 📄

 ## Support 💬

@@ -76,7 +76,7 @@ def serve(
    from open_webui.env import UVICORN_WORKERS  # Import the workers setting

    uvicorn.run(
-        open_webui.main.app,
+        "open_webui.main:app",
        host=host,
        port=port,
        forwarded_allow_ips="*",
@@ -109,54 +109,7 @@ if os.path.exists(f"{DATA_DIR}/config.json"):

 DEFAULT_CONFIG = {
    "version": 0,
-    "ui": {
-        "default_locale": "",
-        "prompt_suggestions": [
-            {
-                "title": [
-                    "Help me study",
-                    "vocabulary for a college entrance exam",
-                ],
-                "content": "Help me study vocabulary: write a sentence for me to fill in the blank, and I'll try to pick the correct option.",
-            },
-            {
-                "title": [
-                    "Give me ideas",
-                    "for what to do with my kids' art",
-                ],
-                "content": "What are 5 creative things I could do with my kids' art? I don't want to throw them away, but it's also so much clutter.",
-            },
-            {
-                "title": ["Tell me a fun fact", "about the Roman Empire"],
-                "content": "Tell me a random fun fact about the Roman Empire",
-            },
-            {
-                "title": [
-                    "Show me a code snippet",
-                    "of a website's sticky header",
-                ],
-                "content": "Show me a code snippet of a website's sticky header in CSS and JavaScript.",
-            },
-            {
-                "title": [
-                    "Explain options trading",
-                    "if I'm familiar with buying and selling stocks",
-                ],
-                "content": "Explain options trading in simple terms if I'm familiar with buying and selling stocks.",
-            },
-            {
-                "title": ["Overcome procrastination", "give me tips"],
-                "content": "Could you start by asking me about instances when I procrastinate the most and then give me some suggestions to overcome it?",
-            },
-            {
-                "title": [
-                    "Grammar check",
-                    "rewrite it for better readability ",
-                ],
-                "content": 'Check the following sentence for grammar and clarity: "[sentence]". Rewrite it for better readability while maintaining its original meaning.',
-            },
-        ],
-    },
+    "ui": {},
 }


@@ -509,6 +462,19 @@ ENABLE_OAUTH_GROUP_MANAGEMENT = PersistentConfig(
    os.environ.get("ENABLE_OAUTH_GROUP_MANAGEMENT", "False").lower() == "true",
 )

+ENABLE_OAUTH_GROUP_CREATION = PersistentConfig(
+    "ENABLE_OAUTH_GROUP_CREATION",
+    "oauth.enable_group_creation",
+    os.environ.get("ENABLE_OAUTH_GROUP_CREATION", "False").lower() == "true",
+)
+
+
+OAUTH_BLOCKED_GROUPS = PersistentConfig(
+    "OAUTH_BLOCKED_GROUPS",
+    "oauth.blocked_groups",
+    os.environ.get("OAUTH_BLOCKED_GROUPS", "[]"),
+)
+
 OAUTH_ROLES_CLAIM = PersistentConfig(
    "OAUTH_ROLES_CLAIM",
    "oauth.roles_claim",
@@ -539,6 +505,12 @@ OAUTH_ALLOWED_DOMAINS = PersistentConfig(
    ],
 )

+OAUTH_UPDATE_PICTURE_ON_LOGIN = PersistentConfig(
+    "OAUTH_UPDATE_PICTURE_ON_LOGIN",
+    "oauth.update_picture_on_login",
+    os.environ.get("OAUTH_UPDATE_PICTURE_ON_LOGIN", "False").lower() == "true",
+)
+

 def load_oauth_providers():
    OAUTH_PROVIDERS.clear()
@@ -748,9 +720,10 @@ S3_BUCKET_NAME = os.environ.get("S3_BUCKET_NAME", None)
 S3_KEY_PREFIX = os.environ.get("S3_KEY_PREFIX", None)
 S3_ENDPOINT_URL = os.environ.get("S3_ENDPOINT_URL", None)
 S3_USE_ACCELERATE_ENDPOINT = (
-    os.environ.get("S3_USE_ACCELERATE_ENDPOINT", "False").lower() == "true"
+    os.environ.get("S3_USE_ACCELERATE_ENDPOINT", "false").lower() == "true"
 )
 S3_ADDRESSING_STYLE = os.environ.get("S3_ADDRESSING_STYLE", None)
+S3_ENABLE_TAGGING = os.getenv("S3_ENABLE_TAGGING", "false").lower() == "true"

 GCS_BUCKET_NAME = os.environ.get("GCS_BUCKET_NAME", None)
 GOOGLE_APPLICATION_CREDENTIALS_JSON = os.environ.get(
@@ -908,11 +881,19 @@ OPENAI_API_BASE_URL = "https://api.openai.com/v1"
 # TOOL_SERVERS
 ####################################

+try:
+    tool_server_connections = json.loads(
+        os.environ.get("TOOL_SERVER_CONNECTIONS", "[]")
+    )
+except Exception as e:
+    log.exception(f"Error loading TOOL_SERVER_CONNECTIONS: {e}")
+    tool_server_connections = []
+

 TOOL_SERVER_CONNECTIONS = PersistentConfig(
    "TOOL_SERVER_CONNECTIONS",
    "tool_server.connections",
-    [],
+    tool_server_connections,
 )

 ####################################
@@ -952,10 +933,15 @@ DEFAULT_MODELS = PersistentConfig(
    "DEFAULT_MODELS", "ui.default_models", os.environ.get("DEFAULT_MODELS", None)
 )

-DEFAULT_PROMPT_SUGGESTIONS = PersistentConfig(
-    "DEFAULT_PROMPT_SUGGESTIONS",
-    "ui.prompt_suggestions",
-    [
+try:
+    default_prompt_suggestions = json.loads(
+        os.environ.get("DEFAULT_PROMPT_SUGGESTIONS", "[]")
+    )
+except Exception as e:
+    log.exception(f"Error loading DEFAULT_PROMPT_SUGGESTIONS: {e}")
+    default_prompt_suggestions = []
+if default_prompt_suggestions == []:
+    default_prompt_suggestions = [
        {
            "title": ["Help me study", "vocabulary for a college entrance exam"],
            "content": "Help me study vocabulary: write a sentence for me to fill in the blank, and I'll try to pick the correct option.",
@@ -983,7 +969,12 @@ DEFAULT_PROMPT_SUGGESTIONS = PersistentConfig(
            "title": ["Overcome procrastination", "give me tips"],
            "content": "Could you start by asking me about instances when I procrastinate the most and then give me some suggestions to overcome it?",
        },
-    ],
+    ]
+
+DEFAULT_PROMPT_SUGGESTIONS = PersistentConfig(
+    "DEFAULT_PROMPT_SUGGESTIONS",
+    "ui.prompt_suggestions",
+    default_prompt_suggestions,
 )

 MODEL_ORDER_LIST = PersistentConfig(
@@ -998,6 +989,26 @@ DEFAULT_USER_ROLE = PersistentConfig(
    os.getenv("DEFAULT_USER_ROLE", "pending"),
 )

+PENDING_USER_OVERLAY_TITLE = PersistentConfig(
+    "PENDING_USER_OVERLAY_TITLE",
+    "ui.pending_user_overlay_title",
+    os.environ.get("PENDING_USER_OVERLAY_TITLE", ""),
+)
+
+PENDING_USER_OVERLAY_CONTENT = PersistentConfig(
+    "PENDING_USER_OVERLAY_CONTENT",
+    "ui.pending_user_overlay_content",
+    os.environ.get("PENDING_USER_OVERLAY_CONTENT", ""),
+)
+
+
+RESPONSE_WATERMARK = PersistentConfig(
+    "RESPONSE_WATERMARK",
+    "ui.watermark",
+    os.environ.get("RESPONSE_WATERMARK", ""),
+)
+
+
 USER_PERMISSIONS_WORKSPACE_MODELS_ACCESS = (
    os.environ.get("USER_PERMISSIONS_WORKSPACE_MODELS_ACCESS", "False").lower()
    == "true"
@@ -1062,6 +1073,26 @@ USER_PERMISSIONS_CHAT_EDIT = (
    os.environ.get("USER_PERMISSIONS_CHAT_EDIT", "True").lower() == "true"
 )

+USER_PERMISSIONS_CHAT_SHARE = (
+    os.environ.get("USER_PERMISSIONS_CHAT_SHARE", "True").lower() == "true"
+)
+
+USER_PERMISSIONS_CHAT_EXPORT = (
+    os.environ.get("USER_PERMISSIONS_CHAT_EXPORT", "True").lower() == "true"
+)
+
+USER_PERMISSIONS_CHAT_STT = (
+    os.environ.get("USER_PERMISSIONS_CHAT_STT", "True").lower() == "true"
+)
+
+USER_PERMISSIONS_CHAT_TTS = (
+    os.environ.get("USER_PERMISSIONS_CHAT_TTS", "True").lower() == "true"
+)
+
+USER_PERMISSIONS_CHAT_CALL = (
+    os.environ.get("USER_PERMISSIONS_CHAT_CALL", "True").lower() == "true"
+)
+
 USER_PERMISSIONS_CHAT_MULTIPLE_MODELS = (
    os.environ.get("USER_PERMISSIONS_CHAT_MULTIPLE_MODELS", "True").lower() == "true"
 )
@@ -1095,6 +1126,10 @@ USER_PERMISSIONS_FEATURES_CODE_INTERPRETER = (
    == "true"
 )

+USER_PERMISSIONS_FEATURES_NOTES = (
+    os.environ.get("USER_PERMISSIONS_FEATURES_NOTES", "True").lower() == "true"
+)
+

 DEFAULT_USER_PERMISSIONS = {
    "workspace": {
@@ -1114,6 +1149,11 @@ DEFAULT_USER_PERMISSIONS = {
        "file_upload": USER_PERMISSIONS_CHAT_FILE_UPLOAD,
        "delete": USER_PERMISSIONS_CHAT_DELETE,
        "edit": USER_PERMISSIONS_CHAT_EDIT,
+        "share": USER_PERMISSIONS_CHAT_SHARE,
+        "export": USER_PERMISSIONS_CHAT_EXPORT,
+        "stt": USER_PERMISSIONS_CHAT_STT,
+        "tts": USER_PERMISSIONS_CHAT_TTS,
+        "call": USER_PERMISSIONS_CHAT_CALL,
        "multiple_models": USER_PERMISSIONS_CHAT_MULTIPLE_MODELS,
        "temporary": USER_PERMISSIONS_CHAT_TEMPORARY,
        "temporary_enforced": USER_PERMISSIONS_CHAT_TEMPORARY_ENFORCED,
@@ -1123,6 +1163,7 @@ DEFAULT_USER_PERMISSIONS = {
        "web_search": USER_PERMISSIONS_FEATURES_WEB_SEARCH,
        "image_generation": USER_PERMISSIONS_FEATURES_IMAGE_GENERATION,
        "code_interpreter": USER_PERMISSIONS_FEATURES_CODE_INTERPRETER,
+        "notes": USER_PERMISSIONS_FEATURES_NOTES,
    },
 }

@@ -1138,6 +1179,11 @@ ENABLE_CHANNELS = PersistentConfig(
    os.environ.get("ENABLE_CHANNELS", "False").lower() == "true",
 )

+ENABLE_NOTES = PersistentConfig(
+    "ENABLE_NOTES",
+    "notes.enable",
+    os.environ.get("ENABLE_NOTES", "True").lower() == "true",
+)

 ENABLE_EVALUATION_ARENA_MODELS = PersistentConfig(
    "ENABLE_EVALUATION_ARENA_MODELS",
@@ -1188,6 +1234,18 @@ ENABLE_USER_WEBHOOKS = PersistentConfig(
    os.environ.get("ENABLE_USER_WEBHOOKS", "True").lower() == "true",
 )

+# FastAPI / AnyIO settings
+THREAD_POOL_SIZE = os.getenv("THREAD_POOL_SIZE", None)
+
+if THREAD_POOL_SIZE is not None and isinstance(THREAD_POOL_SIZE, str):
+    try:
+        THREAD_POOL_SIZE = int(THREAD_POOL_SIZE)
+    except ValueError:
+        log.warning(
+            f"THREAD_POOL_SIZE is not a valid integer: {THREAD_POOL_SIZE}. Defaulting to None."
+        )
+        THREAD_POOL_SIZE = None
+

 def validate_cors_origins(origins):
    for origin in origins:
@@ -1214,7 +1272,9 @@ def validate_cors_origin(origin):
 # To test CORS_ALLOW_ORIGIN locally, you can set something like
 # CORS_ALLOW_ORIGIN=http://localhost:5173;http://localhost:8080
 # in your .env file depending on your frontend port, 5173 in this case.
-CORS_ALLOW_ORIGIN = os.environ.get("CORS_ALLOW_ORIGIN", "*").split(";")
+CORS_ALLOW_ORIGIN = os.environ.get(
+    "CORS_ALLOW_ORIGIN", "*;http://localhost:5173;http://localhost:8080"
+).split(";")

 if "*" in CORS_ALLOW_ORIGIN:
    log.warning(
@@ -1286,6 +1346,9 @@ Generate a concise, 3-5 word title with an emoji summarizing the chat history.
 - Use emojis that enhance understanding of the topic, but avoid quotation marks or special formatting.
 - Write the title in the chat's primary language; default to English if multilingual.
 - Prioritize accuracy over excessive creativity; keep it clear and simple.
+- Your entire response must consist solely of the JSON object, without any introductory or concluding text.
+- The output must be a single, raw JSON object, without any markdown code fences or other encapsulating text.
+- Ensure no conversational text, affirmations, or explanations precede or follow the raw JSON output, as this will cause direct parsing failure.
 ### Output:
 JSON format: { "title": "your concise title here" }
 ### Examples:
@@ -1628,7 +1691,8 @@ DEFAULT_CODE_INTERPRETER_PROMPT = """
 1. **Code Interpreter**: `<code_interpreter type="code" lang="python"></code_interpreter>`
   - You have access to a Python shell that runs directly in the user's browser, enabling fast execution of code for analysis, calculations, or problem-solving.  Use it in this response.
   - The Python code you write can incorporate a wide array of libraries, handle data manipulation or visualization, perform API calls for web-related tasks, or tackle virtually any computational challenge. Use this flexibility to **think outside the box, craft elegant solutions, and harness Python's full potential**.
-   - To use it, **you must enclose your code within `<code_interpreter type="code" lang="python">` XML tags** and stop right away. If you don't, the code won't execute. Do NOT use triple backticks.
+   - To use it, **you must enclose your code within `<code_interpreter type="code" lang="python">` XML tags** and stop right away. If you don't, the code won't execute. 
+   - When writing code in the code_interpreter XML tag, Do NOT use the triple backticks code block for markdown formatting, example: ```py # python code ``` will cause an error because it is markdown formatting, it is not python code.
   - When coding, **always aim to print meaningful outputs** (e.g., results, tables, summaries, or visuals) to better interpret and verify the findings. Avoid relying on implicit outputs; prioritize explicit and clear print statements so the results are effectively communicated to the user.  
   - After obtaining the printed output, **always provide a concise analysis, interpretation, or next steps to help the user understand the findings or refine the outcome further.**  
   - If the results are unclear, unexpected, or require validation, refine the code and execute it again as needed. Always aim to deliver meaningful insights from the results, iterating if necessary.  
@@ -1675,9 +1739,21 @@ MILVUS_URI = os.environ.get("MILVUS_URI", f"{DATA_DIR}/vector_db/milvus.db")
 MILVUS_DB = os.environ.get("MILVUS_DB", "default")
 MILVUS_TOKEN = os.environ.get("MILVUS_TOKEN", None)

+MILVUS_INDEX_TYPE = os.environ.get("MILVUS_INDEX_TYPE", "HNSW")
+MILVUS_METRIC_TYPE = os.environ.get("MILVUS_METRIC_TYPE", "COSINE")
+MILVUS_HNSW_M = int(os.environ.get("MILVUS_HNSW_M", "16"))
+MILVUS_HNSW_EFCONSTRUCTION = int(os.environ.get("MILVUS_HNSW_EFCONSTRUCTION", "100"))
+MILVUS_IVF_FLAT_NLIST = int(os.environ.get("MILVUS_IVF_FLAT_NLIST", "128"))
+
 # Qdrant
 QDRANT_URI = os.environ.get("QDRANT_URI", None)
 QDRANT_API_KEY = os.environ.get("QDRANT_API_KEY", None)
+QDRANT_ON_DISK = os.environ.get("QDRANT_ON_DISK", "false").lower() == "true"
+QDRANT_PREFER_GRPC = os.environ.get("QDRANT_PREFER_GRPC", "False").lower() == "true"
+QDRANT_GRPC_PORT = int(os.environ.get("QDRANT_GRPC_PORT", "6334"))
+ENABLE_QDRANT_MULTITENANCY_MODE = (
+    os.environ.get("ENABLE_QDRANT_MULTITENANCY_MODE", "false").lower() == "true"
+)

 # OpenSearch
 OPENSEARCH_URI = os.environ.get("OPENSEARCH_URI", "https://localhost:9200")
@@ -1709,6 +1785,14 @@ PGVECTOR_INITIALIZE_MAX_VECTOR_LENGTH = int(
    os.environ.get("PGVECTOR_INITIALIZE_MAX_VECTOR_LENGTH", "1536")
 )

+# Pinecone
+PINECONE_API_KEY = os.environ.get("PINECONE_API_KEY", None)
+PINECONE_ENVIRONMENT = os.environ.get("PINECONE_ENVIRONMENT", None)
+PINECONE_INDEX_NAME = os.getenv("PINECONE_INDEX_NAME", "open-webui-index")
+PINECONE_DIMENSION = int(os.getenv("PINECONE_DIMENSION", 1536))  # or 3072, 1024, 768
+PINECONE_METRIC = os.getenv("PINECONE_METRIC", "cosine")
+PINECONE_CLOUD = os.getenv("PINECONE_CLOUD", "aws")  # or "gcp" or "azure"
+
 ####################################
 # Information Retrieval (RAG)
 ####################################
@@ -1745,6 +1829,18 @@ ONEDRIVE_CLIENT_ID = PersistentConfig(
    os.environ.get("ONEDRIVE_CLIENT_ID", ""),
 )

+ONEDRIVE_SHAREPOINT_URL = PersistentConfig(
+    "ONEDRIVE_SHAREPOINT_URL",
+    "onedrive.sharepoint_url",
+    os.environ.get("ONEDRIVE_SHAREPOINT_URL", ""),
+)
+
+ONEDRIVE_SHAREPOINT_TENANT_ID = PersistentConfig(
+    "ONEDRIVE_SHAREPOINT_TENANT_ID",
+    "onedrive.sharepoint_tenant_id",
+    os.environ.get("ONEDRIVE_SHAREPOINT_TENANT_ID", ""),
+)
+
 # RAG Content Extraction
 CONTENT_EXTRACTION_ENGINE = PersistentConfig(
    "CONTENT_EXTRACTION_ENGINE",
@@ -1752,6 +1848,18 @@ CONTENT_EXTRACTION_ENGINE = PersistentConfig(
    os.environ.get("CONTENT_EXTRACTION_ENGINE", "").lower(),
 )

+EXTERNAL_DOCUMENT_LOADER_URL = PersistentConfig(
+    "EXTERNAL_DOCUMENT_LOADER_URL",
+    "rag.external_document_loader_url",
+    os.environ.get("EXTERNAL_DOCUMENT_LOADER_URL", ""),
+)
+
+EXTERNAL_DOCUMENT_LOADER_API_KEY = PersistentConfig(
+    "EXTERNAL_DOCUMENT_LOADER_API_KEY",
+    "rag.external_document_loader_api_key",
+    os.environ.get("EXTERNAL_DOCUMENT_LOADER_API_KEY", ""),
+)
+
 TIKA_SERVER_URL = PersistentConfig(
    "TIKA_SERVER_URL",
    "rag.tika_server_url",
@@ -1764,6 +1872,24 @@ DOCLING_SERVER_URL = PersistentConfig(
    os.getenv("DOCLING_SERVER_URL", "http://docling:5001"),
 )

+DOCLING_OCR_ENGINE = PersistentConfig(
+    "DOCLING_OCR_ENGINE",
+    "rag.docling_ocr_engine",
+    os.getenv("DOCLING_OCR_ENGINE", "tesseract"),
+)
+
+DOCLING_OCR_LANG = PersistentConfig(
+    "DOCLING_OCR_LANG",
+    "rag.docling_ocr_lang",
+    os.getenv("DOCLING_OCR_LANG", "eng,fra,deu,spa"),
+)
+
+DOCLING_DO_PICTURE_DESCRIPTION = PersistentConfig(
+    "DOCLING_DO_PICTURE_DESCRIPTION",
+    "rag.docling_do_picture_description",
+    os.getenv("DOCLING_DO_PICTURE_DESCRIPTION", "False").lower() == "true",
+)
+
 DOCUMENT_INTELLIGENCE_ENDPOINT = PersistentConfig(
    "DOCUMENT_INTELLIGENCE_ENDPOINT",
    "rag.document_intelligence_endpoint",
@@ -1835,6 +1961,16 @@ RAG_FILE_MAX_SIZE = PersistentConfig(
    ),
 )

+RAG_ALLOWED_FILE_EXTENSIONS = PersistentConfig(
+    "RAG_ALLOWED_FILE_EXTENSIONS",
+    "rag.file.allowed_extensions",
+    [
+        ext.strip()
+        for ext in os.environ.get("RAG_ALLOWED_FILE_EXTENSIONS", "").split(",")
+        if ext.strip()
+    ],
+)
+
 RAG_EMBEDDING_ENGINE = PersistentConfig(
    "RAG_EMBEDDING_ENGINE",
    "rag.embedding_engine",
@@ -1880,6 +2016,12 @@ RAG_EMBEDDING_PREFIX_FIELD_NAME = os.environ.get(
    "RAG_EMBEDDING_PREFIX_FIELD_NAME", None
 )

+RAG_RERANKING_ENGINE = PersistentConfig(
+    "RAG_RERANKING_ENGINE",
+    "rag.reranking_engine",
+    os.environ.get("RAG_RERANKING_ENGINE", ""),
+)
+
 RAG_RERANKING_MODEL = PersistentConfig(
    "RAG_RERANKING_MODEL",
    "rag.reranking_model",
@@ -1888,6 +2030,7 @@ RAG_RERANKING_MODEL = PersistentConfig(
 if RAG_RERANKING_MODEL.value != "":
    log.info(f"Reranking model set: {RAG_RERANKING_MODEL.value}")

+
 RAG_RERANKING_MODEL_AUTO_UPDATE = (
    not OFFLINE_MODE
    and os.environ.get("RAG_RERANKING_MODEL_AUTO_UPDATE", "True").lower() == "true"
@@ -1897,6 +2040,18 @@ RAG_RERANKING_MODEL_TRUST_REMOTE_CODE = (
    os.environ.get("RAG_RERANKING_MODEL_TRUST_REMOTE_CODE", "True").lower() == "true"
 )

+RAG_EXTERNAL_RERANKER_URL = PersistentConfig(
+    "RAG_EXTERNAL_RERANKER_URL",
+    "rag.external_reranker_url",
+    os.environ.get("RAG_EXTERNAL_RERANKER_URL", ""),
+)
+
+RAG_EXTERNAL_RERANKER_API_KEY = PersistentConfig(
+    "RAG_EXTERNAL_RERANKER_API_KEY",
+    "rag.external_reranker_api_key",
+    os.environ.get("RAG_EXTERNAL_RERANKER_API_KEY", ""),
+)
+

 RAG_TEXT_SPLITTER = PersistentConfig(
    "RAG_TEXT_SPLITTER",
@@ -2072,6 +2227,24 @@ SEARXNG_QUERY_URL = PersistentConfig(
    os.getenv("SEARXNG_QUERY_URL", ""),
 )

+YACY_QUERY_URL = PersistentConfig(
+    "YACY_QUERY_URL",
+    "rag.web.search.yacy_query_url",
+    os.getenv("YACY_QUERY_URL", ""),
+)
+
+YACY_USERNAME = PersistentConfig(
+    "YACY_USERNAME",
+    "rag.web.search.yacy_username",
+    os.getenv("YACY_USERNAME", ""),
+)
+
+YACY_PASSWORD = PersistentConfig(
+    "YACY_PASSWORD",
+    "rag.web.search.yacy_password",
+    os.getenv("YACY_PASSWORD", ""),
+)
+
 GOOGLE_PSE_API_KEY = PersistentConfig(
    "GOOGLE_PSE_API_KEY",
    "rag.web.search.google_pse_api_key",
@@ -2236,6 +2409,29 @@ FIRECRAWL_API_BASE_URL = PersistentConfig(
    os.environ.get("FIRECRAWL_API_BASE_URL", "https://api.firecrawl.dev"),
 )

+EXTERNAL_WEB_SEARCH_URL = PersistentConfig(
+    "EXTERNAL_WEB_SEARCH_URL",
+    "rag.web.search.external_web_search_url",
+    os.environ.get("EXTERNAL_WEB_SEARCH_URL", ""),
+)
+
+EXTERNAL_WEB_SEARCH_API_KEY = PersistentConfig(
+    "EXTERNAL_WEB_SEARCH_API_KEY",
+    "rag.web.search.external_web_search_api_key",
+    os.environ.get("EXTERNAL_WEB_SEARCH_API_KEY", ""),
+)
+
+EXTERNAL_WEB_LOADER_URL = PersistentConfig(
+    "EXTERNAL_WEB_LOADER_URL",
+    "rag.web.loader.external_web_loader_url",
+    os.environ.get("EXTERNAL_WEB_LOADER_URL", ""),
+)
+
+EXTERNAL_WEB_LOADER_API_KEY = PersistentConfig(
+    "EXTERNAL_WEB_LOADER_API_KEY",
+    "rag.web.loader.external_web_loader_api_key",
+    os.environ.get("EXTERNAL_WEB_LOADER_API_KEY", ""),
+)

 ####################################
 # Images
@@ -2489,6 +2685,14 @@ WHISPER_MODEL_AUTO_UPDATE = (
    and os.environ.get("WHISPER_MODEL_AUTO_UPDATE", "").lower() == "true"
 )

+WHISPER_VAD_FILTER = PersistentConfig(
+    "WHISPER_VAD_FILTER",
+    "audio.stt.whisper_vad_filter",
+    os.getenv("WHISPER_VAD_FILTER", "False").lower() == "true",
+)
+
+WHISPER_LANGUAGE = os.getenv("WHISPER_LANGUAGE", "").lower() or None
+
 # Add Deepgram configuration
 DEEPGRAM_API_KEY = PersistentConfig(
    "DEEPGRAM_API_KEY",
@@ -2496,6 +2700,7 @@ DEEPGRAM_API_KEY = PersistentConfig(
    os.getenv("DEEPGRAM_API_KEY", ""),
 )

+
 AUDIO_STT_OPENAI_API_BASE_URL = PersistentConfig(
    "AUDIO_STT_OPENAI_API_BASE_URL",
    "audio.stt.openai.api_base_url",
@@ -2538,6 +2743,18 @@ AUDIO_STT_AZURE_LOCALES = PersistentConfig(
    os.getenv("AUDIO_STT_AZURE_LOCALES", ""),
 )

+AUDIO_STT_AZURE_BASE_URL = PersistentConfig(
+    "AUDIO_STT_AZURE_BASE_URL",
+    "audio.stt.azure.base_url",
+    os.getenv("AUDIO_STT_AZURE_BASE_URL", ""),
+)
+
+AUDIO_STT_AZURE_MAX_SPEAKERS = PersistentConfig(
+    "AUDIO_STT_AZURE_MAX_SPEAKERS",
+    "audio.stt.azure.max_speakers",
+    os.getenv("AUDIO_STT_AZURE_MAX_SPEAKERS", ""),
+)
+
 AUDIO_TTS_OPENAI_API_BASE_URL = PersistentConfig(
    "AUDIO_TTS_OPENAI_API_BASE_URL",
    "audio.tts.openai.api_base_url",
@@ -2583,7 +2800,13 @@ AUDIO_TTS_SPLIT_ON = PersistentConfig(
 AUDIO_TTS_AZURE_SPEECH_REGION = PersistentConfig(
    "AUDIO_TTS_AZURE_SPEECH_REGION",
    "audio.tts.azure.speech_region",
-    os.getenv("AUDIO_TTS_AZURE_SPEECH_REGION", "eastus"),
+    os.getenv("AUDIO_TTS_AZURE_SPEECH_REGION", ""),
+)
+
+AUDIO_TTS_AZURE_SPEECH_BASE_URL = PersistentConfig(
+    "AUDIO_TTS_AZURE_SPEECH_BASE_URL",
+    "audio.tts.azure.speech_base_url",
+    os.getenv("AUDIO_TTS_AZURE_SPEECH_BASE_URL", ""),
 )

 AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT = PersistentConfig(
@@ -2667,6 +2890,12 @@ LDAP_CA_CERT_FILE = PersistentConfig(
    os.environ.get("LDAP_CA_CERT_FILE", ""),
 )

+LDAP_VALIDATE_CERT = PersistentConfig(
+    "LDAP_VALIDATE_CERT",
+    "ldap.server.validate_cert",
+    os.environ.get("LDAP_VALIDATE_CERT", "True").lower() == "true",
+)
+
 LDAP_CIPHERS = PersistentConfig(
    "LDAP_CIPHERS", "ldap.server.ciphers", os.environ.get("LDAP_CIPHERS", "ALL")
 )
@@ -354,6 +354,10 @@ BYPASS_MODEL_ACCESS_CONTROL = (
    os.environ.get("BYPASS_MODEL_ACCESS_CONTROL", "False").lower() == "true"
 )

+WEBUI_AUTH_SIGNOUT_REDIRECT_URL = os.environ.get(
+    "WEBUI_AUTH_SIGNOUT_REDIRECT_URL", None
+)
+
 ####################################
 # WEBUI_SECRET_KEY
 ####################################
@@ -409,6 +413,11 @@ else:
    except Exception:
        AIOHTTP_CLIENT_TIMEOUT = 300

+
+AIOHTTP_CLIENT_SESSION_SSL = (
+    os.environ.get("AIOHTTP_CLIENT_SESSION_SSL", "True").lower() == "true"
+)
+
 AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST = os.environ.get(
    "AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST",
    os.environ.get("AIOHTTP_CLIENT_TIMEOUT_OPENAI_MODEL_LIST", "10"),
@@ -437,6 +446,56 @@ else:
    except Exception:
        AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER_DATA = 10

+
+AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL = (
+    os.environ.get("AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL", "True").lower() == "true"
+)
+
+
+####################################
+# SENTENCE TRANSFORMERS
+####################################
+
+
+SENTENCE_TRANSFORMERS_BACKEND = os.environ.get("SENTENCE_TRANSFORMERS_BACKEND", "")
+if SENTENCE_TRANSFORMERS_BACKEND == "":
+    SENTENCE_TRANSFORMERS_BACKEND = "torch"
+
+
+SENTENCE_TRANSFORMERS_MODEL_KWARGS = os.environ.get(
+    "SENTENCE_TRANSFORMERS_MODEL_KWARGS", ""
+)
+if SENTENCE_TRANSFORMERS_MODEL_KWARGS == "":
+    SENTENCE_TRANSFORMERS_MODEL_KWARGS = None
+else:
+    try:
+        SENTENCE_TRANSFORMERS_MODEL_KWARGS = json.loads(
+            SENTENCE_TRANSFORMERS_MODEL_KWARGS
+        )
+    except Exception:
+        SENTENCE_TRANSFORMERS_MODEL_KWARGS = None
+
+
+SENTENCE_TRANSFORMERS_CROSS_ENCODER_BACKEND = os.environ.get(
+    "SENTENCE_TRANSFORMERS_CROSS_ENCODER_BACKEND", ""
+)
+if SENTENCE_TRANSFORMERS_CROSS_ENCODER_BACKEND == "":
+    SENTENCE_TRANSFORMERS_CROSS_ENCODER_BACKEND = "torch"
+
+
+SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS = os.environ.get(
+    "SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS", ""
+)
+if SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS == "":
+    SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS = None
+else:
+    try:
+        SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS = json.loads(
+            SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS
+        )
+    except Exception:
+        SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS = None
+
 ####################################
 # OFFLINE_MODE
 ####################################
@@ -446,6 +505,7 @@ OFFLINE_MODE = os.environ.get("OFFLINE_MODE", "false").lower() == "true"
 if OFFLINE_MODE:
    os.environ["HF_HUB_OFFLINE"] = "1"

+
 ####################################
 # AUDIT LOGGING
 ####################################
@@ -467,6 +527,7 @@ AUDIT_EXCLUDED_PATHS = os.getenv("AUDIT_EXCLUDED_PATHS", "/chats,/chat,/folders"
 AUDIT_EXCLUDED_PATHS = [path.strip() for path in AUDIT_EXCLUDED_PATHS]
 AUDIT_EXCLUDED_PATHS = [path.lstrip("/") for path in AUDIT_EXCLUDED_PATHS]

+
 ####################################
 # OPENTELEMETRY
 ####################################
@@ -17,6 +17,7 @@ from sqlalchemy import text
 from typing import Optional
 from aiocache import cached
 import aiohttp
+import anyio.to_thread
 import requests


@@ -63,6 +64,7 @@ from open_webui.routers import (
    auths,
    channels,
    chats,
+    notes,
    folders,
    configs,
    groups,
@@ -100,11 +102,15 @@ from open_webui.config import (
    # OpenAI
    ENABLE_OPENAI_API,
    ONEDRIVE_CLIENT_ID,
+    ONEDRIVE_SHAREPOINT_URL,
+    ONEDRIVE_SHAREPOINT_TENANT_ID,
    OPENAI_API_BASE_URLS,
    OPENAI_API_KEYS,
    OPENAI_API_CONFIGS,
    # Direct Connections
    ENABLE_DIRECT_CONNECTIONS,
+    # Thread pool size for FastAPI/AnyIO
+    THREAD_POOL_SIZE,
    # Tool Server Configs
    TOOL_SERVER_CONNECTIONS,
    # Code Execution
@@ -151,6 +157,8 @@ from open_webui.config import (
    AUDIO_STT_AZURE_API_KEY,
    AUDIO_STT_AZURE_REGION,
    AUDIO_STT_AZURE_LOCALES,
+    AUDIO_STT_AZURE_BASE_URL,
+    AUDIO_STT_AZURE_MAX_SPEAKERS,
    AUDIO_TTS_API_KEY,
    AUDIO_TTS_ENGINE,
    AUDIO_TTS_MODEL,
@@ -159,6 +167,7 @@ from open_webui.config import (
    AUDIO_TTS_SPLIT_ON,
    AUDIO_TTS_VOICE,
    AUDIO_TTS_AZURE_SPEECH_REGION,
+    AUDIO_TTS_AZURE_SPEECH_BASE_URL,
    AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT,
    PLAYWRIGHT_WS_URL,
    PLAYWRIGHT_TIMEOUT,
@@ -166,6 +175,8 @@ from open_webui.config import (
    FIRECRAWL_API_KEY,
    WEB_LOADER_ENGINE,
    WHISPER_MODEL,
+    WHISPER_VAD_FILTER,
+    WHISPER_LANGUAGE,
    DEEPGRAM_API_KEY,
    WHISPER_MODEL_AUTO_UPDATE,
    WHISPER_MODEL_DIR,
@@ -177,12 +188,16 @@ from open_webui.config import (
    RAG_EMBEDDING_MODEL,
    RAG_EMBEDDING_MODEL_AUTO_UPDATE,
    RAG_EMBEDDING_MODEL_TRUST_REMOTE_CODE,
+    RAG_RERANKING_ENGINE,
    RAG_RERANKING_MODEL,
+    RAG_EXTERNAL_RERANKER_URL,
+    RAG_EXTERNAL_RERANKER_API_KEY,
    RAG_RERANKING_MODEL_AUTO_UPDATE,
    RAG_RERANKING_MODEL_TRUST_REMOTE_CODE,
    RAG_EMBEDDING_ENGINE,
    RAG_EMBEDDING_BATCH_SIZE,
    RAG_RELEVANCE_THRESHOLD,
+    RAG_ALLOWED_FILE_EXTENSIONS,
    RAG_FILE_MAX_COUNT,
    RAG_FILE_MAX_SIZE,
    RAG_OPENAI_API_BASE_URL,
@@ -192,8 +207,13 @@ from open_webui.config import (
    CHUNK_OVERLAP,
    CHUNK_SIZE,
    CONTENT_EXTRACTION_ENGINE,
+    EXTERNAL_DOCUMENT_LOADER_URL,
+    EXTERNAL_DOCUMENT_LOADER_API_KEY,
    TIKA_SERVER_URL,
    DOCLING_SERVER_URL,
+    DOCLING_OCR_ENGINE,
+    DOCLING_OCR_LANG,
+    DOCLING_DO_PICTURE_DESCRIPTION,
    DOCUMENT_INTELLIGENCE_ENDPOINT,
    DOCUMENT_INTELLIGENCE_KEY,
    MISTRAL_OCR_API_KEY,
@@ -218,6 +238,9 @@ from open_webui.config import (
    SERPAPI_API_KEY,
    SERPAPI_ENGINE,
    SEARXNG_QUERY_URL,
+    YACY_QUERY_URL,
+    YACY_USERNAME,
+    YACY_PASSWORD,
    SERPER_API_KEY,
    SERPLY_API_KEY,
    SERPSTACK_API_KEY,
@@ -239,12 +262,18 @@ from open_webui.config import (
    GOOGLE_DRIVE_CLIENT_ID,
    GOOGLE_DRIVE_API_KEY,
    ONEDRIVE_CLIENT_ID,
+    ONEDRIVE_SHAREPOINT_URL,
+    ONEDRIVE_SHAREPOINT_TENANT_ID,
    ENABLE_RAG_HYBRID_SEARCH,
    ENABLE_RAG_LOCAL_WEB_FETCH,
    ENABLE_WEB_LOADER_SSL_VERIFICATION,
    ENABLE_GOOGLE_DRIVE_INTEGRATION,
    ENABLE_ONEDRIVE_INTEGRATION,
    UPLOAD_DIR,
+    EXTERNAL_WEB_SEARCH_URL,
+    EXTERNAL_WEB_SEARCH_API_KEY,
+    EXTERNAL_WEB_LOADER_URL,
+    EXTERNAL_WEB_LOADER_API_KEY,
    # WebUI
    WEBUI_AUTH,
    WEBUI_NAME,
@@ -259,12 +288,15 @@ from open_webui.config import (
    ENABLE_API_KEY_ENDPOINT_RESTRICTIONS,
    API_KEY_ALLOWED_ENDPOINTS,
    ENABLE_CHANNELS,
+    ENABLE_NOTES,
    ENABLE_COMMUNITY_SHARING,
    ENABLE_MESSAGE_RATING,
    ENABLE_USER_WEBHOOKS,
    ENABLE_EVALUATION_ARENA_MODELS,
    USER_PERMISSIONS,
    DEFAULT_USER_ROLE,
+    PENDING_USER_OVERLAY_CONTENT,
+    PENDING_USER_OVERLAY_TITLE,
    DEFAULT_PROMPT_SUGGESTIONS,
    DEFAULT_MODELS,
    DEFAULT_ARENA_MODEL,
@@ -291,6 +323,7 @@ from open_webui.config import (
    LDAP_APP_PASSWORD,
    LDAP_USE_TLS,
    LDAP_CA_CERT_FILE,
+    LDAP_VALIDATE_CERT,
    LDAP_CIPHERS,
    # Misc
    ENV,
@@ -301,6 +334,7 @@ from open_webui.config import (
    DEFAULT_LOCALE,
    OAUTH_PROVIDERS,
    WEBUI_URL,
+    RESPONSE_WATERMARK,
    # Admin
    ENABLE_ADMIN_CHAT_ACCESS,
    ENABLE_ADMIN_EXPORT,
@@ -340,12 +374,14 @@ from open_webui.env import (
    WEBUI_SESSION_COOKIE_SECURE,
    WEBUI_AUTH_TRUSTED_EMAIL_HEADER,
    WEBUI_AUTH_TRUSTED_NAME_HEADER,
+    WEBUI_AUTH_SIGNOUT_REDIRECT_URL,
    ENABLE_WEBSOCKET_SUPPORT,
    BYPASS_MODEL_ACCESS_CONTROL,
    RESET_CONFIG_ON_START,
    OFFLINE_MODE,
    ENABLE_OTEL,
    EXTERNAL_PWA_MANIFEST_URL,
+    AIOHTTP_CLIENT_SESSION_SSL,
 )


@@ -369,6 +405,7 @@ from open_webui.utils.auth import (
    get_admin_user,
    get_verified_user,
 )
+from open_webui.utils.plugin import install_tool_and_function_dependencies
 from open_webui.utils.oauth import OAuthManager
 from open_webui.utils.security_headers import SecurityHeadersMiddleware

@@ -415,7 +452,7 @@ print(
 ╚═════╝ ╚═╝     ╚══════╝╚═╝  ╚═══╝     ╚══╝╚══╝ ╚══════╝╚═════╝  ╚═════╝ ╚═╝


-v{VERSION} - building the best open-source AI user interface.
+v{VERSION} - building the best AI user interface.
 {f"Commit: {WEBUI_BUILD_HASH}" if WEBUI_BUILD_HASH != "dev-build" else ""}
 https://github.com/open-webui/open-webui
 """
@@ -431,7 +468,17 @@ async def lifespan(app: FastAPI):
    if LICENSE_KEY:
        get_license_data(app, LICENSE_KEY)

+    # This should be blocking (sync) so functions are not deactivated on first /get_models calls
+    # when the first user lands on the / route.
+    log.info("Installing external dependencies of functions and tools...")
+    install_tool_and_function_dependencies()
+
+    if THREAD_POOL_SIZE and THREAD_POOL_SIZE > 0:
+        limiter = anyio.to_thread.current_default_thread_limiter()
+        limiter.total_tokens = THREAD_POOL_SIZE
+
    asyncio.create_task(periodic_usage_pool_cleanup())
+
    yield


@@ -535,6 +582,11 @@ app.state.config.DEFAULT_MODELS = DEFAULT_MODELS
 app.state.config.DEFAULT_PROMPT_SUGGESTIONS = DEFAULT_PROMPT_SUGGESTIONS
 app.state.config.DEFAULT_USER_ROLE = DEFAULT_USER_ROLE

+app.state.config.PENDING_USER_OVERLAY_CONTENT = PENDING_USER_OVERLAY_CONTENT
+app.state.config.PENDING_USER_OVERLAY_TITLE = PENDING_USER_OVERLAY_TITLE
+
+app.state.config.RESPONSE_WATERMARK = RESPONSE_WATERMARK
+
 app.state.config.USER_PERMISSIONS = USER_PERMISSIONS
 app.state.config.WEBHOOK_URL = WEBHOOK_URL
 app.state.config.BANNERS = WEBUI_BANNERS
@@ -542,6 +594,7 @@ app.state.config.MODEL_ORDER_LIST = MODEL_ORDER_LIST


 app.state.config.ENABLE_CHANNELS = ENABLE_CHANNELS
+app.state.config.ENABLE_NOTES = ENABLE_NOTES
 app.state.config.ENABLE_COMMUNITY_SHARING = ENABLE_COMMUNITY_SHARING
 app.state.config.ENABLE_MESSAGE_RATING = ENABLE_MESSAGE_RATING
 app.state.config.ENABLE_USER_WEBHOOKS = ENABLE_USER_WEBHOOKS
@@ -570,11 +623,13 @@ app.state.config.LDAP_SEARCH_BASE = LDAP_SEARCH_BASE
 app.state.config.LDAP_SEARCH_FILTERS = LDAP_SEARCH_FILTERS
 app.state.config.LDAP_USE_TLS = LDAP_USE_TLS
 app.state.config.LDAP_CA_CERT_FILE = LDAP_CA_CERT_FILE
+app.state.config.LDAP_VALIDATE_CERT = LDAP_VALIDATE_CERT
 app.state.config.LDAP_CIPHERS = LDAP_CIPHERS


 app.state.AUTH_TRUSTED_EMAIL_HEADER = WEBUI_AUTH_TRUSTED_EMAIL_HEADER
 app.state.AUTH_TRUSTED_NAME_HEADER = WEBUI_AUTH_TRUSTED_NAME_HEADER
+app.state.WEBUI_AUTH_SIGNOUT_REDIRECT_URL = WEBUI_AUTH_SIGNOUT_REDIRECT_URL
 app.state.EXTERNAL_PWA_MANIFEST_URL = EXTERNAL_PWA_MANIFEST_URL

 app.state.USER_COUNT = None
@@ -591,6 +646,7 @@ app.state.FUNCTIONS = {}
 app.state.config.TOP_K = RAG_TOP_K
 app.state.config.TOP_K_RERANKER = RAG_TOP_K_RERANKER
 app.state.config.RELEVANCE_THRESHOLD = RAG_RELEVANCE_THRESHOLD
+app.state.config.ALLOWED_FILE_EXTENSIONS = RAG_ALLOWED_FILE_EXTENSIONS
 app.state.config.FILE_MAX_SIZE = RAG_FILE_MAX_SIZE
 app.state.config.FILE_MAX_COUNT = RAG_FILE_MAX_COUNT

@@ -601,8 +657,13 @@ app.state.config.ENABLE_RAG_HYBRID_SEARCH = ENABLE_RAG_HYBRID_SEARCH
 app.state.config.ENABLE_WEB_LOADER_SSL_VERIFICATION = ENABLE_WEB_LOADER_SSL_VERIFICATION

 app.state.config.CONTENT_EXTRACTION_ENGINE = CONTENT_EXTRACTION_ENGINE
+app.state.config.EXTERNAL_DOCUMENT_LOADER_URL = EXTERNAL_DOCUMENT_LOADER_URL
+app.state.config.EXTERNAL_DOCUMENT_LOADER_API_KEY = EXTERNAL_DOCUMENT_LOADER_API_KEY
 app.state.config.TIKA_SERVER_URL = TIKA_SERVER_URL
 app.state.config.DOCLING_SERVER_URL = DOCLING_SERVER_URL
+app.state.config.DOCLING_OCR_ENGINE = DOCLING_OCR_ENGINE
+app.state.config.DOCLING_OCR_LANG = DOCLING_OCR_LANG
+app.state.config.DOCLING_DO_PICTURE_DESCRIPTION = DOCLING_DO_PICTURE_DESCRIPTION
 app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT = DOCUMENT_INTELLIGENCE_ENDPOINT
 app.state.config.DOCUMENT_INTELLIGENCE_KEY = DOCUMENT_INTELLIGENCE_KEY
 app.state.config.MISTRAL_OCR_API_KEY = MISTRAL_OCR_API_KEY
@@ -616,7 +677,12 @@ app.state.config.CHUNK_OVERLAP = CHUNK_OVERLAP
 app.state.config.RAG_EMBEDDING_ENGINE = RAG_EMBEDDING_ENGINE
 app.state.config.RAG_EMBEDDING_MODEL = RAG_EMBEDDING_MODEL
 app.state.config.RAG_EMBEDDING_BATCH_SIZE = RAG_EMBEDDING_BATCH_SIZE
+
+app.state.config.RAG_RERANKING_ENGINE = RAG_RERANKING_ENGINE
 app.state.config.RAG_RERANKING_MODEL = RAG_RERANKING_MODEL
+app.state.config.RAG_EXTERNAL_RERANKER_URL = RAG_EXTERNAL_RERANKER_URL
+app.state.config.RAG_EXTERNAL_RERANKER_API_KEY = RAG_EXTERNAL_RERANKER_API_KEY
+
 app.state.config.RAG_TEMPLATE = RAG_TEMPLATE

 app.state.config.RAG_OPENAI_API_BASE_URL = RAG_OPENAI_API_BASE_URL
@@ -645,6 +711,9 @@ app.state.config.BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL = (
 app.state.config.ENABLE_GOOGLE_DRIVE_INTEGRATION = ENABLE_GOOGLE_DRIVE_INTEGRATION
 app.state.config.ENABLE_ONEDRIVE_INTEGRATION = ENABLE_ONEDRIVE_INTEGRATION
 app.state.config.SEARXNG_QUERY_URL = SEARXNG_QUERY_URL
+app.state.config.YACY_QUERY_URL = YACY_QUERY_URL
+app.state.config.YACY_USERNAME = YACY_USERNAME
+app.state.config.YACY_PASSWORD = YACY_PASSWORD
 app.state.config.GOOGLE_PSE_API_KEY = GOOGLE_PSE_API_KEY
 app.state.config.GOOGLE_PSE_ENGINE_ID = GOOGLE_PSE_ENGINE_ID
 app.state.config.BRAVE_SEARCH_API_KEY = BRAVE_SEARCH_API_KEY
@@ -667,6 +736,10 @@ app.state.config.EXA_API_KEY = EXA_API_KEY
 app.state.config.PERPLEXITY_API_KEY = PERPLEXITY_API_KEY
 app.state.config.SOUGOU_API_SID = SOUGOU_API_SID
 app.state.config.SOUGOU_API_SK = SOUGOU_API_SK
+app.state.config.EXTERNAL_WEB_SEARCH_URL = EXTERNAL_WEB_SEARCH_URL
+app.state.config.EXTERNAL_WEB_SEARCH_API_KEY = EXTERNAL_WEB_SEARCH_API_KEY
+app.state.config.EXTERNAL_WEB_LOADER_URL = EXTERNAL_WEB_LOADER_URL
+app.state.config.EXTERNAL_WEB_LOADER_API_KEY = EXTERNAL_WEB_LOADER_API_KEY


 app.state.config.PLAYWRIGHT_WS_URL = PLAYWRIGHT_WS_URL
@@ -690,7 +763,10 @@ try:
    )

    app.state.rf = get_rf(
+        app.state.config.RAG_RERANKING_ENGINE,
        app.state.config.RAG_RERANKING_MODEL,
+        app.state.config.RAG_EXTERNAL_RERANKER_URL,
+        app.state.config.RAG_EXTERNAL_RERANKER_API_KEY,
        RAG_RERANKING_MODEL_AUTO_UPDATE,
    )
 except Exception as e:
@@ -789,11 +865,14 @@ app.state.config.STT_ENGINE = AUDIO_STT_ENGINE
 app.state.config.STT_MODEL = AUDIO_STT_MODEL

 app.state.config.WHISPER_MODEL = WHISPER_MODEL
+app.state.config.WHISPER_VAD_FILTER = WHISPER_VAD_FILTER
 app.state.config.DEEPGRAM_API_KEY = DEEPGRAM_API_KEY

 app.state.config.AUDIO_STT_AZURE_API_KEY = AUDIO_STT_AZURE_API_KEY
 app.state.config.AUDIO_STT_AZURE_REGION = AUDIO_STT_AZURE_REGION
 app.state.config.AUDIO_STT_AZURE_LOCALES = AUDIO_STT_AZURE_LOCALES
+app.state.config.AUDIO_STT_AZURE_BASE_URL = AUDIO_STT_AZURE_BASE_URL
+app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS = AUDIO_STT_AZURE_MAX_SPEAKERS

 app.state.config.TTS_OPENAI_API_BASE_URL = AUDIO_TTS_OPENAI_API_BASE_URL
 app.state.config.TTS_OPENAI_API_KEY = AUDIO_TTS_OPENAI_API_KEY
@@ -805,6 +884,7 @@ app.state.config.TTS_SPLIT_ON = AUDIO_TTS_SPLIT_ON


 app.state.config.TTS_AZURE_SPEECH_REGION = AUDIO_TTS_AZURE_SPEECH_REGION
+app.state.config.TTS_AZURE_SPEECH_BASE_URL = AUDIO_TTS_AZURE_SPEECH_BASE_URL
 app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT = AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT


@@ -867,7 +947,8 @@ class RedirectMiddleware(BaseHTTPMiddleware):

            # Check for the specific watch path and the presence of 'v' parameter
            if path.endswith("/watch") and "v" in query_params:
-                video_id = query_params["v"][0]  # Extract the first 'v' parameter
+                # Extract the first 'v' parameter
+                video_id = query_params["v"][0]
                encoded_video_id = urlencode({"youtube": video_id})
                redirect_url = f"/?{encoded_video_id}"
                return RedirectResponse(url=redirect_url)
@@ -953,6 +1034,8 @@ app.include_router(users.router, prefix="/api/v1/users", tags=["users"])

 app.include_router(channels.router, prefix="/api/v1/channels", tags=["channels"])
 app.include_router(chats.router, prefix="/api/v1/chats", tags=["chats"])
+app.include_router(notes.router, prefix="/api/v1/notes", tags=["notes"])
+

 app.include_router(models.router, prefix="/api/v1/models", tags=["models"])
 app.include_router(knowledge.router, prefix="/api/v1/knowledge", tags=["knowledge"])
@@ -1023,14 +1106,19 @@ async def get_models(request: Request, user=Depends(get_verified_user)):
        if "pipeline" in model and model["pipeline"].get("type", None) == "filter":
            continue

-        model_tags = [
-            tag.get("name")
-            for tag in model.get("info", {}).get("meta", {}).get("tags", [])
-        ]
-        tags = [tag.get("name") for tag in model.get("tags", [])]
+        try:
+            model_tags = [
+                tag.get("name")
+                for tag in model.get("info", {}).get("meta", {}).get("tags", [])
+            ]
+            tags = [tag.get("name") for tag in model.get("tags", [])]

-        tags = list(set(model_tags + tags))
-        model["tags"] = [{"name": tag} for tag in tags]
+            tags = list(set(model_tags + tags))
+            model["tags"] = [{"name": tag} for tag in tags]
+        except Exception as e:
+            log.debug(f"Error processing model tags: {e}")
+            model["tags"] = []
+            pass

        models.append(model)

@@ -1098,11 +1186,12 @@ async def chat_completion(
            "chat_id": form_data.pop("chat_id", None),
            "message_id": form_data.pop("id", None),
            "session_id": form_data.pop("session_id", None),
+            "filter_ids": form_data.pop("filter_ids", []),
            "tool_ids": form_data.get("tool_ids", None),
            "tool_servers": form_data.pop("tool_servers", None),
            "files": form_data.get("files", None),
-            "features": form_data.get("features", None),
-            "variables": form_data.get("variables", None),
+            "features": form_data.get("features", {}),
+            "variables": form_data.get("variables", {}),
            "model": model,
            "direct": model_item.get("direct", False),
            **(
@@ -1276,6 +1365,7 @@ async def get_app_config(request: Request):
                {
                    "enable_direct_connections": app.state.config.ENABLE_DIRECT_CONNECTIONS,
                    "enable_channels": app.state.config.ENABLE_CHANNELS,
+                    "enable_notes": app.state.config.ENABLE_NOTES,
                    "enable_web_search": app.state.config.ENABLE_WEB_SEARCH,
                    "enable_code_execution": app.state.config.ENABLE_CODE_EXECUTION,
                    "enable_code_interpreter": app.state.config.ENABLE_CODE_INTERPRETER,
@@ -1320,7 +1410,16 @@ async def get_app_config(request: Request):
                    "client_id": GOOGLE_DRIVE_CLIENT_ID.value,
                    "api_key": GOOGLE_DRIVE_API_KEY.value,
                },
-                "onedrive": {"client_id": ONEDRIVE_CLIENT_ID.value},
+                "onedrive": {
+                    "client_id": ONEDRIVE_CLIENT_ID.value,
+                    "sharepoint_url": ONEDRIVE_SHAREPOINT_URL.value,
+                    "sharepoint_tenant_id": ONEDRIVE_SHAREPOINT_TENANT_ID.value,
+                },
+                "ui": {
+                    "pending_user_overlay_title": app.state.config.PENDING_USER_OVERLAY_TITLE,
+                    "pending_user_overlay_content": app.state.config.PENDING_USER_OVERLAY_CONTENT,
+                    "response_watermark": app.state.config.RESPONSE_WATERMARK,
+                },
                "license_metadata": app.state.LICENSE_METADATA,
                **(
                    {
@@ -1372,7 +1471,8 @@ async def get_app_latest_release_version(user=Depends(get_verified_user)):
        timeout = aiohttp.ClientTimeout(total=1)
        async with aiohttp.ClientSession(timeout=timeout, trust_env=True) as session:
            async with session.get(
-                "https://api.github.com/repos/open-webui/open-webui/releases/latest"
+                "https://api.github.com/repos/open-webui/open-webui/releases/latest",
+                ssl=AIOHTTP_CLIENT_SESSION_SSL,
            ) as response:
                response.raise_for_status()
                data = await response.json()
@@ -1432,7 +1532,7 @@ async def get_manifest_json():
            "start_url": "/",
            "display": "standalone",
            "background_color": "#343541",
-            "orientation": "natural",
+            "orientation": "any",
            "icons": [
                {
                    "src": "/static/logo.png",
@@ -0,0 +1,33 @@
+"""Add note table
+
+Revision ID: 9f0c9cd09105
+Revises: 3781e22d8b01
+Create Date: 2025-05-03 03:00:00.000000
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+revision = "9f0c9cd09105"
+down_revision = "3781e22d8b01"
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    op.create_table(
+        "note",
+        sa.Column("id", sa.Text(), nullable=False, primary_key=True, unique=True),
+        sa.Column("user_id", sa.Text(), nullable=True),
+        sa.Column("title", sa.Text(), nullable=True),
+        sa.Column("data", sa.JSON(), nullable=True),
+        sa.Column("meta", sa.JSON(), nullable=True),
+        sa.Column("access_control", sa.JSON(), nullable=True),
+        sa.Column("created_at", sa.BigInteger(), nullable=True),
+        sa.Column("updated_at", sa.BigInteger(), nullable=True),
+    )
+
+
+def downgrade():
+    op.drop_table("note")
@@ -436,7 +436,7 @@ class ChatTable:

            all_chats = query.all()

-            # result has to be destrctured from sqlalchemy `row` and mapped to a dict since the `ChatModel`is not the returned dataclass.
+            # result has to be destructured from sqlalchemy `row` and mapped to a dict since the `ChatModel`is not the returned dataclass.
            return [
                ChatTitleIdResponse.model_validate(
                    {
@@ -0,0 +1,135 @@
+import json
+import time
+import uuid
+from typing import Optional
+
+from open_webui.internal.db import Base, get_db
+from open_webui.utils.access_control import has_access
+from open_webui.models.users import Users, UserResponse
+
+
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Boolean, Column, String, Text, JSON
+from sqlalchemy import or_, func, select, and_, text
+from sqlalchemy.sql import exists
+
+####################
+# Note DB Schema
+####################
+
+
+class Note(Base):
+    __tablename__ = "note"
+
+    id = Column(Text, primary_key=True)
+    user_id = Column(Text)
+
+    title = Column(Text)
+    data = Column(JSON, nullable=True)
+    meta = Column(JSON, nullable=True)
+
+    access_control = Column(JSON, nullable=True)
+
+    created_at = Column(BigInteger)
+    updated_at = Column(BigInteger)
+
+
+class NoteModel(BaseModel):
+    model_config = ConfigDict(from_attributes=True)
+
+    id: str
+    user_id: str
+
+    title: str
+    data: Optional[dict] = None
+    meta: Optional[dict] = None
+
+    access_control: Optional[dict] = None
+
+    created_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
+
+
+####################
+# Forms
+####################
+
+
+class NoteForm(BaseModel):
+    title: str
+    data: Optional[dict] = None
+    meta: Optional[dict] = None
+    access_control: Optional[dict] = None
+
+
+class NoteUserResponse(NoteModel):
+    user: Optional[UserResponse] = None
+
+
+class NoteTable:
+    def insert_new_note(
+        self,
+        form_data: NoteForm,
+        user_id: str,
+    ) -> Optional[NoteModel]:
+        with get_db() as db:
+            note = NoteModel(
+                **{
+                    "id": str(uuid.uuid4()),
+                    "user_id": user_id,
+                    **form_data.model_dump(),
+                    "created_at": int(time.time_ns()),
+                    "updated_at": int(time.time_ns()),
+                }
+            )
+
+            new_note = Note(**note.model_dump())
+
+            db.add(new_note)
+            db.commit()
+            return note
+
+    def get_notes(self) -> list[NoteModel]:
+        with get_db() as db:
+            notes = db.query(Note).order_by(Note.updated_at.desc()).all()
+            return [NoteModel.model_validate(note) for note in notes]
+
+    def get_notes_by_user_id(
+        self, user_id: str, permission: str = "write"
+    ) -> list[NoteModel]:
+        notes = self.get_notes()
+        return [
+            note
+            for note in notes
+            if note.user_id == user_id
+            or has_access(user_id, permission, note.access_control)
+        ]
+
+    def get_note_by_id(self, id: str) -> Optional[NoteModel]:
+        with get_db() as db:
+            note = db.query(Note).filter(Note.id == id).first()
+            return NoteModel.model_validate(note) if note else None
+
+    def update_note_by_id(self, id: str, form_data: NoteForm) -> Optional[NoteModel]:
+        with get_db() as db:
+            note = db.query(Note).filter(Note.id == id).first()
+            if not note:
+                return None
+
+            note.title = form_data.title
+            note.data = form_data.data
+            note.meta = form_data.meta
+            note.access_control = form_data.access_control
+            note.updated_at = int(time.time_ns())
+
+            db.commit()
+            return NoteModel.model_validate(note) if note else None
+
+    def delete_note_by_id(self, id: str):
+        with get_db() as db:
+            db.query(Note).filter(Note.id == id).delete()
+            db.commit()
+            return True
+
+
+Notes = NoteTable()
@@ -10,6 +10,8 @@ from open_webui.models.groups import Groups

 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Column, String, Text
+from sqlalchemy import or_
+

 ####################
 # User DB Schema
@@ -67,6 +69,11 @@ class UserModel(BaseModel):
 ####################


+class UserListResponse(BaseModel):
+    users: list[UserModel]
+    total: int
+
+
 class UserResponse(BaseModel):
    id: str
    name: str
@@ -160,11 +167,63 @@ class UsersTable:
            return None

    def get_users(
-        self, skip: Optional[int] = None, limit: Optional[int] = None
-    ) -> list[UserModel]:
+        self,
+        filter: Optional[dict] = None,
+        skip: Optional[int] = None,
+        limit: Optional[int] = None,
+    ) -> UserListResponse:
        with get_db() as db:
+            query = db.query(User)

-            query = db.query(User).order_by(User.created_at.desc())
+            if filter:
+                query_key = filter.get("query")
+                if query_key:
+                    query = query.filter(
+                        or_(
+                            User.name.ilike(f"%{query_key}%"),
+                            User.email.ilike(f"%{query_key}%"),
+                        )
+                    )
+
+                order_by = filter.get("order_by")
+                direction = filter.get("direction")
+
+                if order_by == "name":
+                    if direction == "asc":
+                        query = query.order_by(User.name.asc())
+                    else:
+                        query = query.order_by(User.name.desc())
+                elif order_by == "email":
+                    if direction == "asc":
+                        query = query.order_by(User.email.asc())
+                    else:
+                        query = query.order_by(User.email.desc())
+
+                elif order_by == "created_at":
+                    if direction == "asc":
+                        query = query.order_by(User.created_at.asc())
+                    else:
+                        query = query.order_by(User.created_at.desc())
+
+                elif order_by == "last_active_at":
+                    if direction == "asc":
+                        query = query.order_by(User.last_active_at.asc())
+                    else:
+                        query = query.order_by(User.last_active_at.desc())
+
+                elif order_by == "updated_at":
+                    if direction == "asc":
+                        query = query.order_by(User.updated_at.asc())
+                    else:
+                        query = query.order_by(User.updated_at.desc())
+                elif order_by == "role":
+                    if direction == "asc":
+                        query = query.order_by(User.role.asc())
+                    else:
+                        query = query.order_by(User.role.desc())
+
+            else:
+                query = query.order_by(User.created_at.desc())

            if skip:
                query = query.offset(skip)
@@ -172,8 +231,10 @@ class UsersTable:
                query = query.limit(limit)

            users = query.all()
-
-            return [UserModel.model_validate(user) for user in users]
+            return {
+                "users": [UserModel.model_validate(user) for user in users],
+                "total": db.query(User).count(),
+            }

    def get_users_by_user_ids(self, user_ids: list[str]) -> list[UserModel]:
        with get_db() as db:
@@ -330,5 +391,13 @@ class UsersTable:
            users = db.query(User).filter(User.id.in_(user_ids)).all()
            return [user.id for user in users]

+    def get_super_admin_user(self) -> Optional[UserModel]:
+        with get_db() as db:
+            user = db.query(User).filter_by(role="admin").first()
+            if user:
+                return UserModel.model_validate(user)
+            else:
+                return None
+

 Users = UsersTable()
@@ -0,0 +1,58 @@
+import requests
+import logging
+from typing import Iterator, List, Union
+
+from langchain_core.document_loaders import BaseLoader
+from langchain_core.documents import Document
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+class ExternalDocumentLoader(BaseLoader):
+    def __init__(
+        self,
+        file_path,
+        url: str,
+        api_key: str,
+        mime_type=None,
+        **kwargs,
+    ) -> None:
+        self.url = url
+        self.api_key = api_key
+
+        self.file_path = file_path
+        self.mime_type = mime_type
+
+    def load(self) -> list[Document]:
+        with open(self.file_path, "rb") as f:
+            data = f.read()
+
+        headers = {}
+        if self.mime_type is not None:
+            headers["Content-Type"] = self.mime_type
+
+        if self.api_key is not None:
+            headers["Authorization"] = f"Bearer {self.api_key}"
+
+        url = self.url
+        if url.endswith("/"):
+            url = url[:-1]
+
+        r = requests.put(f"{url}/process", data=data, headers=headers)
+
+        if r.ok:
+            res = r.json()
+
+            if res:
+                return [
+                    Document(
+                        page_content=res.get("page_content"),
+                        metadata=res.get("metadata"),
+                    )
+                ]
+            else:
+                raise Exception("Error loading document: No content returned")
+        else:
+            raise Exception(f"Error loading document: {r.status_code} {r.text}")
@@ -0,0 +1,53 @@
+import requests
+import logging
+from typing import Iterator, List, Union
+
+from langchain_core.document_loaders import BaseLoader
+from langchain_core.documents import Document
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+class ExternalWebLoader(BaseLoader):
+    def __init__(
+        self,
+        web_paths: Union[str, List[str]],
+        external_url: str,
+        external_api_key: str,
+        continue_on_failure: bool = True,
+        **kwargs,
+    ) -> None:
+        self.external_url = external_url
+        self.external_api_key = external_api_key
+        self.urls = web_paths if isinstance(web_paths, list) else [web_paths]
+        self.continue_on_failure = continue_on_failure
+
+    def lazy_load(self) -> Iterator[Document]:
+        batch_size = 20
+        for i in range(0, len(self.urls), batch_size):
+            urls = self.urls[i : i + batch_size]
+            try:
+                response = requests.post(
+                    self.external_url,
+                    headers={
+                        "User-Agent": "Open WebUI (https://github.com/open-webui/open-webui) External Web Loader",
+                        "Authorization": f"Bearer {self.external_api_key}",
+                    },
+                    json={
+                        "urls": urls,
+                    },
+                )
+                response.raise_for_status()
+                results = response.json()
+                for result in results:
+                    yield Document(
+                        page_content=result.get("page_content", ""),
+                        metadata=result.get("metadata", {}),
+                    )
+            except Exception as e:
+                if self.continue_on_failure:
+                    log.error(f"Error extracting content from batch {urls}: {e}")
+                else:
+                    raise e
@@ -21,6 +21,8 @@ from langchain_community.document_loaders import (
 )
 from langchain_core.documents import Document

+
+from open_webui.retrieval.loaders.external_document import ExternalDocumentLoader
 from open_webui.retrieval.loaders.mistral import MistralLoader

 from open_webui.env import SRC_LOG_LEVELS, GLOBAL_LOG_LEVEL
@@ -85,11 +87,13 @@ known_source_ext = [


 class TikaLoader:
-    def __init__(self, url, file_path, mime_type=None):
+    def __init__(self, url, file_path, mime_type=None, extract_images=None):
        self.url = url
        self.file_path = file_path
        self.mime_type = mime_type

+        self.extract_images = extract_images
+
    def load(self) -> list[Document]:
        with open(self.file_path, "rb") as f:
            data = f.read()
@@ -99,6 +103,9 @@ class TikaLoader:
        else:
            headers = {}

+        if self.extract_images == True:
+            headers["X-Tika-PDFextractInlineImages"] = "true"
+
        endpoint = self.url
        if not endpoint.endswith("/"):
            endpoint += "/"
@@ -121,11 +128,13 @@ class TikaLoader:


 class DoclingLoader:
-    def __init__(self, url, file_path=None, mime_type=None):
+    def __init__(self, url, file_path=None, mime_type=None, params=None):
        self.url = url.rstrip("/")
        self.file_path = file_path
        self.mime_type = mime_type

+        self.params = params or {}
+
    def load(self) -> list[Document]:
        with open(self.file_path, "rb") as f:
            files = {
@@ -141,6 +150,20 @@ class DoclingLoader:
                "table_mode": "accurate",
            }

+            if self.params:
+                if self.params.get("do_picture_classification"):
+                    params["do_picture_classification"] = self.params.get(
+                        "do_picture_classification"
+                    )
+
+                if self.params.get("ocr_engine") and self.params.get("ocr_lang"):
+                    params["ocr_engine"] = self.params.get("ocr_engine")
+                    params["ocr_lang"] = [
+                        lang.strip()
+                        for lang in self.params.get("ocr_lang").split(",")
+                        if lang.strip()
+                    ]
+
            endpoint = f"{self.url}/v1alpha/convert/file"
            r = requests.post(endpoint, files=files, data=params)

@@ -192,6 +215,17 @@ class Loader:
    def _get_loader(self, filename: str, file_content_type: str, file_path: str):
        file_ext = filename.split(".")[-1].lower()

+        if (
+            self.engine == "external"
+            and self.kwargs.get("EXTERNAL_DOCUMENT_LOADER_URL")
+            and self.kwargs.get("EXTERNAL_DOCUMENT_LOADER_API_KEY")
+        ):
+            loader = ExternalDocumentLoader(
+                file_path=file_path,
+                url=self.kwargs.get("EXTERNAL_DOCUMENT_LOADER_URL"),
+                api_key=self.kwargs.get("EXTERNAL_DOCUMENT_LOADER_API_KEY"),
+                mime_type=file_content_type,
+            )
        if self.engine == "tika" and self.kwargs.get("TIKA_SERVER_URL"):
            if self._is_text_file(file_ext, file_content_type):
                loader = TextLoader(file_path, autodetect_encoding=True)
@@ -200,6 +234,7 @@ class Loader:
                    url=self.kwargs.get("TIKA_SERVER_URL"),
                    file_path=file_path,
                    mime_type=file_content_type,
+                    extract_images=self.kwargs.get("PDF_EXTRACT_IMAGES"),
                )
        elif self.engine == "docling" and self.kwargs.get("DOCLING_SERVER_URL"):
            if self._is_text_file(file_ext, file_content_type):
@@ -209,6 +244,13 @@ class Loader:
                    url=self.kwargs.get("DOCLING_SERVER_URL"),
                    file_path=file_path,
                    mime_type=file_content_type,
+                    params={
+                        "ocr_engine": self.kwargs.get("DOCLING_OCR_ENGINE"),
+                        "ocr_lang": self.kwargs.get("DOCLING_OCR_LANG"),
+                        "do_picture_classification": self.kwargs.get(
+                            "DOCLING_DO_PICTURE_DESCRIPTION"
+                        ),
+                    },
                )
        elif (
            self.engine == "document_intelligence"
@@ -240,6 +282,15 @@ class Loader:
            loader = MistralLoader(
                api_key=self.kwargs.get("MISTRAL_OCR_API_KEY"), file_path=file_path
            )
+        elif (
+            self.engine == "external"
+            and self.kwargs.get("MISTRAL_OCR_API_KEY") != ""
+            and file_ext
+            in ["pdf"]  # Mistral OCR currently only supports PDF and images
+        ):
+            loader = MistralLoader(
+                api_key=self.kwargs.get("MISTRAL_OCR_API_KEY"), file_path=file_path
+            )
        else:
            if file_ext == "pdf":
                loader = PyPDFLoader(
@@ -62,12 +62,17 @@ class YoutubeLoader:
        _video_id = _parse_video_id(video_id)
        self.video_id = _video_id if _video_id is not None else video_id
        self._metadata = {"source": video_id}
-        self.language = language
        self.proxy_url = proxy_url
+
+        # Ensure language is a list
        if isinstance(language, str):
            self.language = [language]
        else:
-            self.language = language
+            self.language = list(language)
+
+        # Add English as fallback if not already in the list
+        if "en" not in self.language:
+            self.language.append("en")

    def load(self) -> List[Document]:
        """Load YouTube transcripts into `Document` objects."""
@@ -101,17 +106,31 @@ class YoutubeLoader:
            log.exception("Loading YouTube transcript failed")
            return []

-        try:
-            transcript = transcript_list.find_transcript(self.language)
-        except NoTranscriptFound:
-            transcript = transcript_list.find_transcript(["en"])
+        # Try each language in order of priority
+        for lang in self.language:
+            try:
+                transcript = transcript_list.find_transcript([lang])
+                log.debug(f"Found transcript for language '{lang}'")
+                transcript_pieces: List[Dict[str, Any]] = transcript.fetch()
+                transcript_text = " ".join(
+                    map(
+                        lambda transcript_piece: transcript_piece.text.strip(" "),
+                        transcript_pieces,
+                    )
+                )
+                return [Document(page_content=transcript_text, metadata=self._metadata)]
+            except NoTranscriptFound:
+                log.debug(f"No transcript found for language '{lang}'")
+                continue
+            except Exception as e:
+                log.info(f"Error finding transcript for language '{lang}'")
+                raise e

-        transcript_pieces: List[Dict[str, Any]] = transcript.fetch()
-
-        transcript = " ".join(
-            map(
-                lambda transcript_piece: transcript_piece.text.strip(" "),
-                transcript_pieces,
-            )
+        # If we get here, all languages failed
+        languages_tried = ", ".join(self.language)
+        log.warning(
+            f"No transcript found for any of the specified languages: {languages_tried}. Verify if the video has transcripts, add more languages if needed."
+        )
+        raise NoTranscriptFound(
+            f"No transcript found for any supported language. Verify if the video has transcripts, add more languages if needed."
        )
-        return [Document(page_content=transcript, metadata=self._metadata)]
@@ -0,0 +1,58 @@
+import logging
+import requests
+from typing import Optional, List, Tuple
+
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+class ExternalReranker:
+    def __init__(
+        self,
+        api_key: str,
+        url: str = "http://localhost:8080/v1/rerank",
+        model: str = "reranker",
+    ):
+        self.api_key = api_key
+        self.url = url
+        self.model = model
+
+    def predict(self, sentences: List[Tuple[str, str]]) -> Optional[List[float]]:
+        query = sentences[0][0]
+        docs = [i[1] for i in sentences]
+
+        payload = {
+            "model": self.model,
+            "query": query,
+            "documents": docs,
+            "top_n": len(docs),
+        }
+
+        try:
+            log.info(f"ExternalReranker:predict:model {self.model}")
+            log.info(f"ExternalReranker:predict:query {query}")
+
+            r = requests.post(
+                f"{self.url}",
+                headers={
+                    "Content-Type": "application/json",
+                    "Authorization": f"Bearer {self.api_key}",
+                },
+                json=payload,
+            )
+
+            r.raise_for_status()
+            data = r.json()
+
+            if "results" in data:
+                sorted_results = sorted(data["results"], key=lambda x: x["index"])
+                return [result["relevance_score"] for result in sorted_results]
+            else:
+                log.error("No results found in external reranking response")
+                return None
+
+        except Exception as e:
+            log.exception(f"Error in external reranking: {e}")
+            return None
@@ -12,7 +12,7 @@ from langchain_community.retrievers import BM25Retriever
 from langchain_core.documents import Document

 from open_webui.config import VECTOR_DB
-from open_webui.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.retrieval.vector.factory import VECTOR_DB_CLIENT

 from open_webui.models.users import UserModel
 from open_webui.models.files import Files
@@ -207,7 +207,7 @@ def merge_and_sort_query_results(query_results: list[dict], k: int) -> dict:

        for distance, document, metadata in zip(distances, documents, metadatas):
            if isinstance(document, str):
-                doc_hash = hashlib.md5(
+                doc_hash = hashlib.sha256(
                    document.encode()
                ).hexdigest()  # Compute a hash for uniqueness

@@ -260,23 +260,47 @@ def query_collection(
    k: int,
 ) -> dict:
    results = []
-    for query in queries:
-        log.debug(f"query_collection:query {query}")
-        query_embedding = embedding_function(query, prefix=RAG_EMBEDDING_QUERY_PREFIX)
-        for collection_name in collection_names:
+    error = False
+
+    def process_query_collection(collection_name, query_embedding):
+        try:
            if collection_name:
-                try:
-                    result = query_doc(
-                        collection_name=collection_name,
-                        k=k,
-                        query_embedding=query_embedding,
-                    )
-                    if result is not None:
-                        results.append(result.model_dump())
-                except Exception as e:
-                    log.exception(f"Error when querying the collection: {e}")
-            else:
-                pass
+                result = query_doc(
+                    collection_name=collection_name,
+                    k=k,
+                    query_embedding=query_embedding,
+                )
+                if result is not None:
+                    return result.model_dump(), None
+            return None, None
+        except Exception as e:
+            log.exception(f"Error when querying the collection: {e}")
+            return None, e
+
+    # Generate all query embeddings (in one call)
+    query_embeddings = embedding_function(queries, prefix=RAG_EMBEDDING_QUERY_PREFIX)
+    log.debug(
+        f"query_collection: processing {len(queries)} queries across {len(collection_names)} collections"
+    )
+
+    with ThreadPoolExecutor() as executor:
+        future_results = []
+        for query_embedding in query_embeddings:
+            for collection_name in collection_names:
+                result = executor.submit(
+                    process_query_collection, collection_name, query_embedding
+                )
+                future_results.append(result)
+        task_results = [future.result() for future in future_results]
+
+    for result, err in task_results:
+        if err is not None:
+            error = True
+        elif result is not None:
+            results.append(result)
+
+    if error and not results:
+        log.warning("All collection queries failed. No results returned.")

    return merge_and_sort_query_results(results, k=k)

@@ -794,7 +818,9 @@ class RerankCompressor(BaseDocumentCompressor):
            )
            scores = util.cos_sim(query_embedding, document_embedding)[0]

-        docs_with_scores = list(zip(documents, scores.tolist()))
+        docs_with_scores = list(
+            zip(documents, scores.tolist() if not isinstance(scores, list) else scores)
+        )
        if self.r_score:
            docs_with_scores = [
                (d, s) for d, s in docs_with_scores if s >= self.r_score
@@ -1,26 +0,0 @@
-from open_webui.config import VECTOR_DB
-
-if VECTOR_DB == "milvus":
-    from open_webui.retrieval.vector.dbs.milvus import MilvusClient
-
-    VECTOR_DB_CLIENT = MilvusClient()
-elif VECTOR_DB == "qdrant":
-    from open_webui.retrieval.vector.dbs.qdrant import QdrantClient
-
-    VECTOR_DB_CLIENT = QdrantClient()
-elif VECTOR_DB == "opensearch":
-    from open_webui.retrieval.vector.dbs.opensearch import OpenSearchClient
-
-    VECTOR_DB_CLIENT = OpenSearchClient()
-elif VECTOR_DB == "pgvector":
-    from open_webui.retrieval.vector.dbs.pgvector import PgvectorClient
-
-    VECTOR_DB_CLIENT = PgvectorClient()
-elif VECTOR_DB == "elasticsearch":
-    from open_webui.retrieval.vector.dbs.elasticsearch import ElasticsearchClient
-
-    VECTOR_DB_CLIENT = ElasticsearchClient()
-else:
-    from open_webui.retrieval.vector.dbs.chroma import ChromaClient
-
-    VECTOR_DB_CLIENT = ChromaClient()
@@ -5,7 +5,12 @@ from chromadb.utils.batch_utils import create_batches

 from typing import Optional

-from open_webui.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
 from open_webui.config import (
    CHROMA_DATA_PATH,
    CHROMA_HTTP_HOST,
@@ -23,7 +28,7 @@ log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])


-class ChromaClient:
+class ChromaClient(VectorDBBase):
    def __init__(self):
        settings_dict = {
            "allow_reset": True,
@@ -2,7 +2,12 @@ from elasticsearch import Elasticsearch, BadRequestError
 from typing import Optional
 import ssl
 from elasticsearch.helpers import bulk, scan
-from open_webui.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
 from open_webui.config import (
    ELASTICSEARCH_URL,
    ELASTICSEARCH_CA_CERTS,
@@ -15,7 +20,7 @@ from open_webui.config import (
 )


-class ElasticsearchClient:
+class ElasticsearchClient(VectorDBBase):
    """
    Important:
    in order to reduce the number of indexes and since the embedding vector length is fixed, we avoid creating
@@ -3,12 +3,21 @@ from pymilvus import FieldSchema, DataType
 import json
 import logging
 from typing import Optional
-
-from open_webui.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
 from open_webui.config import (
    MILVUS_URI,
    MILVUS_DB,
    MILVUS_TOKEN,
+    MILVUS_INDEX_TYPE,
+    MILVUS_METRIC_TYPE,
+    MILVUS_HNSW_M,
+    MILVUS_HNSW_EFCONSTRUCTION,
+    MILVUS_IVF_FLAT_NLIST,
 )
 from open_webui.env import SRC_LOG_LEVELS

@@ -16,7 +25,7 @@ log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])


-class MilvusClient:
+class MilvusClient(VectorDBBase):
    def __init__(self):
        self.collection_prefix = "open_webui"
        if MILVUS_TOKEN is None:
@@ -28,7 +37,6 @@ class MilvusClient:
        ids = []
        documents = []
        metadatas = []
-
        for match in result:
            _ids = []
            _documents = []
@@ -37,11 +45,9 @@ class MilvusClient:
                _ids.append(item.get("id"))
                _documents.append(item.get("data", {}).get("text"))
                _metadatas.append(item.get("metadata"))
-
            ids.append(_ids)
            documents.append(_documents)
            metadatas.append(_metadatas)
-
        return GetResult(
            **{
                "ids": ids,
@@ -55,13 +61,11 @@ class MilvusClient:
        distances = []
        documents = []
        metadatas = []
-
        for match in result:
            _ids = []
            _distances = []
            _documents = []
            _metadatas = []
-
            for item in match:
                _ids.append(item.get("id"))
                # normalize milvus score from [-1, 1] to [0, 1] range
@@ -70,12 +74,10 @@ class MilvusClient:
                _distances.append(_dist)
                _documents.append(item.get("entity", {}).get("data", {}).get("text"))
                _metadatas.append(item.get("entity", {}).get("metadata"))
-
            ids.append(_ids)
            distances.append(_distances)
            documents.append(_documents)
            metadatas.append(_metadatas)
-
        return SearchResult(
            **{
                "ids": ids,
@@ -108,11 +110,39 @@ class MilvusClient:
        )

        index_params = self.client.prepare_index_params()
+
+        # Use configurations from config.py
+        index_type = MILVUS_INDEX_TYPE.upper()
+        metric_type = MILVUS_METRIC_TYPE.upper()
+
+        log.info(f"Using Milvus index type: {index_type}, metric type: {metric_type}")
+
+        index_creation_params = {}
+        if index_type == "HNSW":
+            index_creation_params = {
+                "M": MILVUS_HNSW_M,
+                "efConstruction": MILVUS_HNSW_EFCONSTRUCTION,
+            }
+            log.info(f"HNSW params: {index_creation_params}")
+        elif index_type == "IVF_FLAT":
+            index_creation_params = {"nlist": MILVUS_IVF_FLAT_NLIST}
+            log.info(f"IVF_FLAT params: {index_creation_params}")
+        elif index_type in ["FLAT", "AUTOINDEX"]:
+            log.info(f"Using {index_type} index with no specific build-time params.")
+        else:
+            log.warning(
+                f"Unsupported MILVUS_INDEX_TYPE: '{index_type}'. "
+                f"Supported types: HNSW, IVF_FLAT, FLAT, AUTOINDEX. "
+                f"Milvus will use its default for the collection if this type is not directly supported for index creation."
+            )
+            # For unsupported types, pass the type directly to Milvus; it might handle it or use a default.
+            # If Milvus errors out, the user needs to correct the MILVUS_INDEX_TYPE env var.
+
        index_params.add_index(
            field_name="vector",
-            index_type="HNSW",
-            metric_type="COSINE",
-            params={"M": 16, "efConstruction": 100},
+            index_type=index_type,
+            metric_type=metric_type,
+            params=index_creation_params,
        )

        self.client.create_collection(
@@ -120,6 +150,9 @@ class MilvusClient:
            schema=schema,
            index_params=index_params,
        )
+        log.info(
+            f"Successfully created collection '{self.collection_prefix}_{collection_name}' with index type '{index_type}' and metric '{metric_type}'."
+        )

    def has_collection(self, collection_name: str) -> bool:
        # Check if the collection exists based on the collection name.
@@ -140,84 +173,113 @@ class MilvusClient:
    ) -> Optional[SearchResult]:
        # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
        collection_name = collection_name.replace("-", "_")
+        # For some index types like IVF_FLAT, search params like nprobe can be set.
+        # Example: search_params = {"nprobe": 10} if using IVF_FLAT
+        # For simplicity, not adding configurable search_params here, but could be extended.
        result = self.client.search(
            collection_name=f"{self.collection_prefix}_{collection_name}",
            data=vectors,
            limit=limit,
            output_fields=["data", "metadata"],
+            # search_params=search_params # Potentially add later if needed
        )
-
        return self._result_to_search_result(result)

    def query(self, collection_name: str, filter: dict, limit: Optional[int] = None):
        # Construct the filter string for querying
        collection_name = collection_name.replace("-", "_")
        if not self.has_collection(collection_name):
+            log.warning(
+                f"Query attempted on non-existent collection: {self.collection_prefix}_{collection_name}"
+            )
            return None
-
        filter_string = " && ".join(
            [
                f'metadata["{key}"] == {json.dumps(value)}'
                for key, value in filter.items()
            ]
        )
-
        max_limit = 16383  # The maximum number of records per request
        all_results = []
-
        if limit is None:
-            limit = float("inf")  # Use infinity as a placeholder for no limit
+            # Milvus default limit for query if not specified is 16384, but docs mention iteration.
+            # Let's set a practical high number if "all" is intended, or handle true pagination.
+            # For now, if limit is None, we'll fetch in batches up to a very large number.
+            # This part could be refined based on expected use cases for "get all".
+            # For this function signature, None implies "as many as possible" up to Milvus limits.
+            limit = (
+                16384 * 10
+            )  # A large number to signify fetching many, will be capped by actual data or max_limit per call.
+            log.info(
+                f"Limit not specified for query, fetching up to {limit} results in batches."
+            )

        # Initialize offset and remaining to handle pagination
        offset = 0
        remaining = limit

        try:
+            log.info(
+                f"Querying collection {self.collection_prefix}_{collection_name} with filter: '{filter_string}', limit: {limit}"
+            )
            # Loop until there are no more items to fetch or the desired limit is reached
            while remaining > 0:
-                log.info(f"remaining: {remaining}")
                current_fetch = min(
-                    max_limit, remaining
-                )  # Determine how many items to fetch in this iteration
+                    max_limit, remaining if isinstance(remaining, int) else max_limit
+                )
+                log.debug(
+                    f"Querying with offset: {offset}, current_fetch: {current_fetch}"
+                )

                results = self.client.query(
                    collection_name=f"{self.collection_prefix}_{collection_name}",
                    filter=filter_string,
-                    output_fields=["*"],
+                    output_fields=[
+                        "id",
+                        "data",
+                        "metadata",
+                    ],  # Explicitly list needed fields. Vector not usually needed in query.
                    limit=current_fetch,
                    offset=offset,
                )

                if not results:
+                    log.debug("No more results from query.")
                    break

                all_results.extend(results)
                results_count = len(results)
-                remaining -= (
-                    results_count  # Decrease remaining by the number of items fetched
-                )
+                log.debug(f"Fetched {results_count} results in this batch.")
+
+                if isinstance(remaining, int):
+                    remaining -= results_count
+
                offset += results_count

-                # Break the loop if the results returned are less than the requested fetch count
+                # Break the loop if the results returned are less than the requested fetch count (means end of data)
                if results_count < current_fetch:
+                    log.debug(
+                        "Fetched less than requested, assuming end of results for this query."
+                    )
                    break

-            log.debug(all_results)
+            log.info(f"Total results from query: {len(all_results)}")
            return self._result_to_get_result([all_results])
        except Exception as e:
            log.exception(
-                f"Error querying collection {collection_name} with limit {limit}: {e}"
+                f"Error querying collection {self.collection_prefix}_{collection_name} with filter '{filter_string}' and limit {limit}: {e}"
            )
            return None

    def get(self, collection_name: str) -> Optional[GetResult]:
-        # Get all the items in the collection.
+        # Get all the items in the collection. This can be very resource-intensive for large collections.
        collection_name = collection_name.replace("-", "_")
-        result = self.client.query(
-            collection_name=f"{self.collection_prefix}_{collection_name}",
-            filter='id != ""',
+        log.warning(
+            f"Fetching ALL items from collection '{self.collection_prefix}_{collection_name}'. This might be slow for large collections."
        )
-        return self._result_to_get_result([result])
+        # Using query with a trivial filter to get all items.
+        # This will use the paginated query logic.
+        return self.query(collection_name=collection_name, filter={}, limit=None)

    def insert(self, collection_name: str, items: list[VectorItem]):
        # Insert the items into the collection, if the collection does not exist, it will be created.
@@ -225,10 +287,23 @@ class MilvusClient:
        if not self.client.has_collection(
            collection_name=f"{self.collection_prefix}_{collection_name}"
        ):
+            log.info(
+                f"Collection {self.collection_prefix}_{collection_name} does not exist. Creating now."
+            )
+            if not items:
+                log.error(
+                    f"Cannot create collection {self.collection_prefix}_{collection_name} without items to determine dimension."
+                )
+                raise ValueError(
+                    "Cannot create Milvus collection without items to determine vector dimension."
+                )
            self._create_collection(
                collection_name=collection_name, dimension=len(items[0]["vector"])
            )

+        log.info(
+            f"Inserting {len(items)} items into collection {self.collection_prefix}_{collection_name}."
+        )
        return self.client.insert(
            collection_name=f"{self.collection_prefix}_{collection_name}",
            data=[
@@ -248,10 +323,23 @@ class MilvusClient:
        if not self.client.has_collection(
            collection_name=f"{self.collection_prefix}_{collection_name}"
        ):
+            log.info(
+                f"Collection {self.collection_prefix}_{collection_name} does not exist for upsert. Creating now."
+            )
+            if not items:
+                log.error(
+                    f"Cannot create collection {self.collection_prefix}_{collection_name} for upsert without items to determine dimension."
+                )
+                raise ValueError(
+                    "Cannot create Milvus collection for upsert without items to determine vector dimension."
+                )
            self._create_collection(
                collection_name=collection_name, dimension=len(items[0]["vector"])
            )

+        log.info(
+            f"Upserting {len(items)} items into collection {self.collection_prefix}_{collection_name}."
+        )
        return self.client.upsert(
            collection_name=f"{self.collection_prefix}_{collection_name}",
            data=[
@@ -271,30 +359,55 @@ class MilvusClient:
        ids: Optional[list[str]] = None,
        filter: Optional[dict] = None,
    ):
-        # Delete the items from the collection based on the ids.
+        # Delete the items from the collection based on the ids or filter.
        collection_name = collection_name.replace("-", "_")
+        if not self.has_collection(collection_name):
+            log.warning(
+                f"Delete attempted on non-existent collection: {self.collection_prefix}_{collection_name}"
+            )
+            return None
+
        if ids:
+            log.info(
+                f"Deleting items by IDs from {self.collection_prefix}_{collection_name}. IDs: {ids}"
+            )
            return self.client.delete(
                collection_name=f"{self.collection_prefix}_{collection_name}",
                ids=ids,
            )
        elif filter:
-            # Convert the filter dictionary to a string using JSON_CONTAINS.
            filter_string = " && ".join(
                [
                    f'metadata["{key}"] == {json.dumps(value)}'
                    for key, value in filter.items()
                ]
            )
-
+            log.info(
+                f"Deleting items by filter from {self.collection_prefix}_{collection_name}. Filter: {filter_string}"
+            )
            return self.client.delete(
                collection_name=f"{self.collection_prefix}_{collection_name}",
                filter=filter_string,
            )
+        else:
+            log.warning(
+                f"Delete operation on {self.collection_prefix}_{collection_name} called without IDs or filter. No action taken."
+            )
+            return None

    def reset(self):
-        # Resets the database. This will delete all collections and item entries.
+        # Resets the database. This will delete all collections and item entries that match the prefix.
+        log.warning(
+            f"Resetting Milvus: Deleting all collections with prefix '{self.collection_prefix}'."
+        )
        collection_names = self.client.list_collections()
-        for collection_name in collection_names:
-            if collection_name.startswith(self.collection_prefix):
-                self.client.drop_collection(collection_name=collection_name)
+        deleted_collections = []
+        for collection_name_full in collection_names:
+            if collection_name_full.startswith(self.collection_prefix):
+                try:
+                    self.client.drop_collection(collection_name=collection_name_full)
+                    deleted_collections.append(collection_name_full)
+                    log.info(f"Deleted collection: {collection_name_full}")
+                except Exception as e:
+                    log.error(f"Error deleting collection {collection_name_full}: {e}")
+        log.info(f"Milvus reset complete. Deleted collections: {deleted_collections}")
@@ -2,7 +2,12 @@ from opensearchpy import OpenSearch
 from opensearchpy.helpers import bulk
 from typing import Optional

-from open_webui.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
 from open_webui.config import (
    OPENSEARCH_URI,
    OPENSEARCH_SSL,
@@ -12,7 +17,7 @@ from open_webui.config import (
 )


-class OpenSearchClient:
+class OpenSearchClient(VectorDBBase):
    def __init__(self):
        self.index_prefix = "open_webui"
        self.client = OpenSearch(
@@ -22,7 +22,12 @@ from pgvector.sqlalchemy import Vector
 from sqlalchemy.ext.mutable import MutableDict
 from sqlalchemy.exc import NoSuchTableError

-from open_webui.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
 from open_webui.config import PGVECTOR_DB_URL, PGVECTOR_INITIALIZE_MAX_VECTOR_LENGTH

 from open_webui.env import SRC_LOG_LEVELS
@@ -44,7 +49,7 @@ class DocumentChunk(Base):
    vmetadata = Column(MutableDict.as_mutable(JSONB), nullable=True)


-class PgvectorClient:
+class PgvectorClient(VectorDBBase):
    def __init__(self) -> None:

        # if no pgvector uri, use the existing database connection
@@ -136,9 +141,8 @@ class PgvectorClient:
            # Pad the vector with zeros
            vector += [0.0] * (VECTOR_LENGTH - current_length)
        elif current_length > VECTOR_LENGTH:
-            raise Exception(
-                f"Vector length {current_length} not supported. Max length must be <= {VECTOR_LENGTH}"
-            )
+            # Truncate the vector to VECTOR_LENGTH
+            vector = vector[:VECTOR_LENGTH]
        return vector

    def insert(self, collection_name: str, items: List[VectorItem]) -> None:
@@ -0,0 +1,533 @@
+from typing import Optional, List, Dict, Any, Union
+import logging
+import time  # for measuring elapsed time
+from pinecone import ServerlessSpec
+
+import asyncio  # for async upserts
+import functools  # for partial binding in async tasks
+
+import concurrent.futures  # for parallel batch upserts
+from pinecone.grpc import PineconeGRPC  # use gRPC client for faster upserts
+
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
+from open_webui.config import (
+    PINECONE_API_KEY,
+    PINECONE_ENVIRONMENT,
+    PINECONE_INDEX_NAME,
+    PINECONE_DIMENSION,
+    PINECONE_METRIC,
+    PINECONE_CLOUD,
+)
+from open_webui.env import SRC_LOG_LEVELS
+
+NO_LIMIT = 10000  # Reasonable limit to avoid overwhelming the system
+BATCH_SIZE = 100  # Recommended batch size for Pinecone operations
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+class PineconeClient(VectorDBBase):
+    def __init__(self):
+        self.collection_prefix = "open-webui"
+
+        # Validate required configuration
+        self._validate_config()
+
+        # Store configuration values
+        self.api_key = PINECONE_API_KEY
+        self.environment = PINECONE_ENVIRONMENT
+        self.index_name = PINECONE_INDEX_NAME
+        self.dimension = PINECONE_DIMENSION
+        self.metric = PINECONE_METRIC
+        self.cloud = PINECONE_CLOUD
+
+        # Initialize Pinecone gRPC client for improved performance
+        self.client = PineconeGRPC(
+            api_key=self.api_key, environment=self.environment, cloud=self.cloud
+        )
+
+        # Persistent executor for batch operations
+        self._executor = concurrent.futures.ThreadPoolExecutor(max_workers=5)
+
+        # Create index if it doesn't exist
+        self._initialize_index()
+
+    def _validate_config(self) -> None:
+        """Validate that all required configuration variables are set."""
+        missing_vars = []
+        if not PINECONE_API_KEY:
+            missing_vars.append("PINECONE_API_KEY")
+        if not PINECONE_ENVIRONMENT:
+            missing_vars.append("PINECONE_ENVIRONMENT")
+        if not PINECONE_INDEX_NAME:
+            missing_vars.append("PINECONE_INDEX_NAME")
+        if not PINECONE_DIMENSION:
+            missing_vars.append("PINECONE_DIMENSION")
+        if not PINECONE_CLOUD:
+            missing_vars.append("PINECONE_CLOUD")
+
+        if missing_vars:
+            raise ValueError(
+                f"Required configuration missing: {', '.join(missing_vars)}"
+            )
+
+    def _initialize_index(self) -> None:
+        """Initialize the Pinecone index."""
+        try:
+            # Check if index exists
+            if self.index_name not in self.client.list_indexes().names():
+                log.info(f"Creating Pinecone index '{self.index_name}'...")
+                self.client.create_index(
+                    name=self.index_name,
+                    dimension=self.dimension,
+                    metric=self.metric,
+                    spec=ServerlessSpec(cloud=self.cloud, region=self.environment),
+                )
+                log.info(f"Successfully created Pinecone index '{self.index_name}'")
+            else:
+                log.info(f"Using existing Pinecone index '{self.index_name}'")
+
+            # Connect to the index
+            self.index = self.client.Index(self.index_name)
+
+        except Exception as e:
+            log.error(f"Failed to initialize Pinecone index: {e}")
+            raise RuntimeError(f"Failed to initialize Pinecone index: {e}")
+
+    def _create_points(
+        self, items: List[VectorItem], collection_name_with_prefix: str
+    ) -> List[Dict[str, Any]]:
+        """Convert VectorItem objects to Pinecone point format."""
+        points = []
+        for item in items:
+            # Start with any existing metadata or an empty dict
+            metadata = item.get("metadata", {}).copy() if item.get("metadata") else {}
+
+            # Add text to metadata if available
+            if "text" in item:
+                metadata["text"] = item["text"]
+
+            # Always add collection_name to metadata for filtering
+            metadata["collection_name"] = collection_name_with_prefix
+
+            point = {
+                "id": item["id"],
+                "values": item["vector"],
+                "metadata": metadata,
+            }
+            points.append(point)
+        return points
+
+    def _get_collection_name_with_prefix(self, collection_name: str) -> str:
+        """Get the collection name with prefix."""
+        return f"{self.collection_prefix}_{collection_name}"
+
+    def _normalize_distance(self, score: float) -> float:
+        """Normalize distance score based on the metric used."""
+        if self.metric.lower() == "cosine":
+            # Cosine similarity ranges from -1 to 1, normalize to 0 to 1
+            return (score + 1.0) / 2.0
+        elif self.metric.lower() in ["euclidean", "dotproduct"]:
+            # These are already suitable for ranking (smaller is better for Euclidean)
+            return score
+        else:
+            # For other metrics, use as is
+            return score
+
+    def _result_to_get_result(self, matches: list) -> GetResult:
+        """Convert Pinecone matches to GetResult format."""
+        ids = []
+        documents = []
+        metadatas = []
+
+        for match in matches:
+            metadata = match.get("metadata", {})
+            ids.append(match["id"])
+            documents.append(metadata.get("text", ""))
+            metadatas.append(metadata)
+
+        return GetResult(
+            **{
+                "ids": [ids],
+                "documents": [documents],
+                "metadatas": [metadatas],
+            }
+        )
+
+    def has_collection(self, collection_name: str) -> bool:
+        """Check if a collection exists by searching for at least one item."""
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+
+        try:
+            # Search for at least 1 item with this collection name in metadata
+            response = self.index.query(
+                vector=[0.0] * self.dimension,  # dummy vector
+                top_k=1,
+                filter={"collection_name": collection_name_with_prefix},
+                include_metadata=False,
+            )
+            return len(response.matches) > 0
+        except Exception as e:
+            log.exception(
+                f"Error checking collection '{collection_name_with_prefix}': {e}"
+            )
+            return False
+
+    def delete_collection(self, collection_name: str) -> None:
+        """Delete a collection by removing all vectors with the collection name in metadata."""
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+        try:
+            self.index.delete(filter={"collection_name": collection_name_with_prefix})
+            log.info(
+                f"Collection '{collection_name_with_prefix}' deleted (all vectors removed)."
+            )
+        except Exception as e:
+            log.warning(
+                f"Failed to delete collection '{collection_name_with_prefix}': {e}"
+            )
+            raise
+
+    def insert(self, collection_name: str, items: List[VectorItem]) -> None:
+        """Insert vectors into a collection."""
+        if not items:
+            log.warning("No items to insert")
+            return
+
+        start_time = time.time()
+
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+        points = self._create_points(items, collection_name_with_prefix)
+
+        # Parallelize batch inserts for performance
+        executor = self._executor
+        futures = []
+        for i in range(0, len(points), BATCH_SIZE):
+            batch = points[i : i + BATCH_SIZE]
+            futures.append(executor.submit(self.index.upsert, vectors=batch))
+        for future in concurrent.futures.as_completed(futures):
+            try:
+                future.result()
+            except Exception as e:
+                log.error(f"Error inserting batch: {e}")
+                raise
+        elapsed = time.time() - start_time
+        log.debug(f"Insert of {len(points)} vectors took {elapsed:.2f} seconds")
+        log.info(
+            f"Successfully inserted {len(points)} vectors in parallel batches into '{collection_name_with_prefix}'"
+        )
+
+    def upsert(self, collection_name: str, items: List[VectorItem]) -> None:
+        """Upsert (insert or update) vectors into a collection."""
+        if not items:
+            log.warning("No items to upsert")
+            return
+
+        start_time = time.time()
+
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+        points = self._create_points(items, collection_name_with_prefix)
+
+        # Parallelize batch upserts for performance
+        executor = self._executor
+        futures = []
+        for i in range(0, len(points), BATCH_SIZE):
+            batch = points[i : i + BATCH_SIZE]
+            futures.append(executor.submit(self.index.upsert, vectors=batch))
+        for future in concurrent.futures.as_completed(futures):
+            try:
+                future.result()
+            except Exception as e:
+                log.error(f"Error upserting batch: {e}")
+                raise
+        elapsed = time.time() - start_time
+        log.debug(f"Upsert of {len(points)} vectors took {elapsed:.2f} seconds")
+        log.info(
+            f"Successfully upserted {len(points)} vectors in parallel batches into '{collection_name_with_prefix}'"
+        )
+
+    async def insert_async(self, collection_name: str, items: List[VectorItem]) -> None:
+        """Async version of insert using asyncio and run_in_executor for improved performance."""
+        if not items:
+            log.warning("No items to insert")
+            return
+
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+        points = self._create_points(items, collection_name_with_prefix)
+
+        # Create batches
+        batches = [
+            points[i : i + BATCH_SIZE] for i in range(0, len(points), BATCH_SIZE)
+        ]
+        loop = asyncio.get_event_loop()
+        tasks = [
+            loop.run_in_executor(
+                None, functools.partial(self.index.upsert, vectors=batch)
+            )
+            for batch in batches
+        ]
+        results = await asyncio.gather(*tasks, return_exceptions=True)
+        for result in results:
+            if isinstance(result, Exception):
+                log.error(f"Error in async insert batch: {result}")
+                raise result
+        log.info(
+            f"Successfully async inserted {len(points)} vectors in batches into '{collection_name_with_prefix}'"
+        )
+
+    async def upsert_async(self, collection_name: str, items: List[VectorItem]) -> None:
+        """Async version of upsert using asyncio and run_in_executor for improved performance."""
+        if not items:
+            log.warning("No items to upsert")
+            return
+
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+        points = self._create_points(items, collection_name_with_prefix)
+
+        # Create batches
+        batches = [
+            points[i : i + BATCH_SIZE] for i in range(0, len(points), BATCH_SIZE)
+        ]
+        loop = asyncio.get_event_loop()
+        tasks = [
+            loop.run_in_executor(
+                None, functools.partial(self.index.upsert, vectors=batch)
+            )
+            for batch in batches
+        ]
+        results = await asyncio.gather(*tasks, return_exceptions=True)
+        for result in results:
+            if isinstance(result, Exception):
+                log.error(f"Error in async upsert batch: {result}")
+                raise result
+        log.info(
+            f"Successfully async upserted {len(points)} vectors in batches into '{collection_name_with_prefix}'"
+        )
+
+    def streaming_upsert(self, collection_name: str, items: List[VectorItem]) -> None:
+        """Perform a streaming upsert over gRPC for performance testing."""
+        if not items:
+            log.warning("No items to upsert via streaming")
+            return
+
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+        points = self._create_points(items, collection_name_with_prefix)
+
+        # Open a streaming upsert channel
+        stream = self.index.streaming_upsert()
+        try:
+            for point in points:
+                # send each point over the stream
+                stream.send(point)
+            # close the stream to finalize
+            stream.close()
+            log.info(
+                f"Successfully streamed upsert of {len(points)} vectors into '{collection_name_with_prefix}'"
+            )
+        except Exception as e:
+            log.error(f"Error during streaming upsert: {e}")
+            raise
+
+    def search(
+        self, collection_name: str, vectors: List[List[Union[float, int]]], limit: int
+    ) -> Optional[SearchResult]:
+        """Search for similar vectors in a collection."""
+        if not vectors or not vectors[0]:
+            log.warning("No vectors provided for search")
+            return None
+
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+
+        if limit is None or limit <= 0:
+            limit = NO_LIMIT
+
+        try:
+            # Search using the first vector (assuming this is the intended behavior)
+            query_vector = vectors[0]
+
+            # Perform the search
+            query_response = self.index.query(
+                vector=query_vector,
+                top_k=limit,
+                include_metadata=True,
+                filter={"collection_name": collection_name_with_prefix},
+            )
+
+            if not query_response.matches:
+                # Return empty result if no matches
+                return SearchResult(
+                    ids=[[]],
+                    documents=[[]],
+                    metadatas=[[]],
+                    distances=[[]],
+                )
+
+            # Convert to GetResult format
+            get_result = self._result_to_get_result(query_response.matches)
+
+            # Calculate normalized distances based on metric
+            distances = [
+                [
+                    self._normalize_distance(match.score)
+                    for match in query_response.matches
+                ]
+            ]
+
+            return SearchResult(
+                ids=get_result.ids,
+                documents=get_result.documents,
+                metadatas=get_result.metadatas,
+                distances=distances,
+            )
+        except Exception as e:
+            log.error(f"Error searching in '{collection_name_with_prefix}': {e}")
+            return None
+
+    def query(
+        self, collection_name: str, filter: Dict, limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        """Query vectors by metadata filter."""
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+
+        if limit is None or limit <= 0:
+            limit = NO_LIMIT
+
+        try:
+            # Create a zero vector for the dimension as Pinecone requires a vector
+            zero_vector = [0.0] * self.dimension
+
+            # Combine user filter with collection_name
+            pinecone_filter = {"collection_name": collection_name_with_prefix}
+            if filter:
+                pinecone_filter.update(filter)
+
+            # Perform metadata-only query
+            query_response = self.index.query(
+                vector=zero_vector,
+                filter=pinecone_filter,
+                top_k=limit,
+                include_metadata=True,
+            )
+
+            return self._result_to_get_result(query_response.matches)
+
+        except Exception as e:
+            log.error(f"Error querying collection '{collection_name}': {e}")
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        """Get all vectors in a collection."""
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+
+        try:
+            # Use a zero vector for fetching all entries
+            zero_vector = [0.0] * self.dimension
+
+            # Add filter to only get vectors for this collection
+            query_response = self.index.query(
+                vector=zero_vector,
+                top_k=NO_LIMIT,
+                include_metadata=True,
+                filter={"collection_name": collection_name_with_prefix},
+            )
+
+            return self._result_to_get_result(query_response.matches)
+
+        except Exception as e:
+            log.error(f"Error getting collection '{collection_name}': {e}")
+            return None
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[List[str]] = None,
+        filter: Optional[Dict] = None,
+    ) -> None:
+        """Delete vectors by IDs or filter."""
+        collection_name_with_prefix = self._get_collection_name_with_prefix(
+            collection_name
+        )
+
+        try:
+            if ids:
+                # Delete by IDs (in batches for large deletions)
+                for i in range(0, len(ids), BATCH_SIZE):
+                    batch_ids = ids[i : i + BATCH_SIZE]
+                    # Note: When deleting by ID, we can't filter by collection_name
+                    # This is a limitation of Pinecone - be careful with ID uniqueness
+                    self.index.delete(ids=batch_ids)
+                    log.debug(
+                        f"Deleted batch of {len(batch_ids)} vectors by ID from '{collection_name_with_prefix}'"
+                    )
+                log.info(
+                    f"Successfully deleted {len(ids)} vectors by ID from '{collection_name_with_prefix}'"
+                )
+
+            elif filter:
+                # Combine user filter with collection_name
+                pinecone_filter = {"collection_name": collection_name_with_prefix}
+                if filter:
+                    pinecone_filter.update(filter)
+                # Delete by metadata filter
+                self.index.delete(filter=pinecone_filter)
+                log.info(
+                    f"Successfully deleted vectors by filter from '{collection_name_with_prefix}'"
+                )
+
+            else:
+                log.warning("No ids or filter provided for delete operation")
+
+        except Exception as e:
+            log.error(f"Error deleting from collection '{collection_name}': {e}")
+            raise
+
+    def reset(self) -> None:
+        """Reset the database by deleting all collections."""
+        try:
+            self.index.delete(delete_all=True)
+            log.info("All vectors successfully deleted from the index.")
+        except Exception as e:
+            log.error(f"Failed to reset Pinecone index: {e}")
+            raise
+
+    def close(self):
+        """Shut down the gRPC channel and thread pool."""
+        try:
+            self.client.close()
+            log.info("Pinecone gRPC channel closed.")
+        except Exception as e:
+            log.warning(f"Failed to close Pinecone gRPC channel: {e}")
+        self._executor.shutdown(wait=True)
+
+    def __enter__(self):
+        """Enter context manager."""
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        """Exit context manager, ensuring resources are cleaned up."""
+        self.close()
@@ -1,12 +1,24 @@
 from typing import Optional
 import logging
+from urllib.parse import urlparse

 from qdrant_client import QdrantClient as Qclient
 from qdrant_client.http.models import PointStruct
 from qdrant_client.models import models

-from open_webui.retrieval.vector.main import VectorItem, SearchResult, GetResult
-from open_webui.config import QDRANT_URI, QDRANT_API_KEY
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
+from open_webui.config import (
+    QDRANT_URI,
+    QDRANT_API_KEY,
+    QDRANT_ON_DISK,
+    QDRANT_GRPC_PORT,
+    QDRANT_PREFER_GRPC,
+)
 from open_webui.env import SRC_LOG_LEVELS

 NO_LIMIT = 999999999
@@ -15,16 +27,34 @@ log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])


-class QdrantClient:
+class QdrantClient(VectorDBBase):
    def __init__(self):
        self.collection_prefix = "open-webui"
        self.QDRANT_URI = QDRANT_URI
        self.QDRANT_API_KEY = QDRANT_API_KEY
-        self.client = (
-            Qclient(url=self.QDRANT_URI, api_key=self.QDRANT_API_KEY)
-            if self.QDRANT_URI
-            else None
-        )
+        self.QDRANT_ON_DISK = QDRANT_ON_DISK
+        self.PREFER_GRPC = QDRANT_PREFER_GRPC
+        self.GRPC_PORT = QDRANT_GRPC_PORT
+
+        if not self.QDRANT_URI:
+            self.client = None
+            return
+
+        # Unified handling for either scheme
+        parsed = urlparse(self.QDRANT_URI)
+        host = parsed.hostname or self.QDRANT_URI
+        http_port = parsed.port or 6333  # default REST port
+
+        if self.PREFER_GRPC:
+            self.client = Qclient(
+                host=host,
+                port=http_port,
+                grpc_port=self.GRPC_PORT,
+                prefer_grpc=self.PREFER_GRPC,
+                api_key=self.QDRANT_API_KEY,
+            )
+        else:
+            self.client = Qclient(url=self.QDRANT_URI, api_key=self.QDRANT_API_KEY)

    def _result_to_get_result(self, points) -> GetResult:
        ids = []
@@ -50,7 +80,9 @@ class QdrantClient:
        self.client.create_collection(
            collection_name=collection_name_with_prefix,
            vectors_config=models.VectorParams(
-                size=dimension, distance=models.Distance.COSINE
+                size=dimension,
+                distance=models.Distance.COSINE,
+                on_disk=self.QDRANT_ON_DISK,
            ),
        )

@@ -0,0 +1,712 @@
+import logging
+from typing import Optional, Tuple
+from urllib.parse import urlparse
+
+import grpc
+from open_webui.config import (
+    QDRANT_API_KEY,
+    QDRANT_GRPC_PORT,
+    QDRANT_ON_DISK,
+    QDRANT_PREFER_GRPC,
+    QDRANT_URI,
+)
+from open_webui.env import SRC_LOG_LEVELS
+from open_webui.retrieval.vector.main import (
+    GetResult,
+    SearchResult,
+    VectorDBBase,
+    VectorItem,
+)
+from qdrant_client import QdrantClient as Qclient
+from qdrant_client.http.exceptions import UnexpectedResponse
+from qdrant_client.http.models import PointStruct
+from qdrant_client.models import models
+
+NO_LIMIT = 999999999
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+class QdrantClient(VectorDBBase):
+    def __init__(self):
+        self.collection_prefix = "open-webui"
+        self.QDRANT_URI = QDRANT_URI
+        self.QDRANT_API_KEY = QDRANT_API_KEY
+        self.QDRANT_ON_DISK = QDRANT_ON_DISK
+        self.PREFER_GRPC = QDRANT_PREFER_GRPC
+        self.GRPC_PORT = QDRANT_GRPC_PORT
+
+        if not self.QDRANT_URI:
+            self.client = None
+            return
+
+        # Unified handling for either scheme
+        parsed = urlparse(self.QDRANT_URI)
+        host = parsed.hostname or self.QDRANT_URI
+        http_port = parsed.port or 6333  # default REST port
+
+        if self.PREFER_GRPC:
+            self.client = Qclient(
+                host=host,
+                port=http_port,
+                grpc_port=self.GRPC_PORT,
+                prefer_grpc=self.PREFER_GRPC,
+                api_key=self.QDRANT_API_KEY,
+            )
+        else:
+            self.client = Qclient(url=self.QDRANT_URI, api_key=self.QDRANT_API_KEY)
+
+        # Main collection types for multi-tenancy
+        self.MEMORY_COLLECTION = f"{self.collection_prefix}_memories"
+        self.KNOWLEDGE_COLLECTION = f"{self.collection_prefix}_knowledge"
+        self.FILE_COLLECTION = f"{self.collection_prefix}_files"
+        self.WEB_SEARCH_COLLECTION = f"{self.collection_prefix}_web-search"
+        self.HASH_BASED_COLLECTION = f"{self.collection_prefix}_hash-based"
+
+    def _result_to_get_result(self, points) -> GetResult:
+        ids = []
+        documents = []
+        metadatas = []
+
+        for point in points:
+            payload = point.payload
+            ids.append(point.id)
+            documents.append(payload["text"])
+            metadatas.append(payload["metadata"])
+
+        return GetResult(
+            **{
+                "ids": [ids],
+                "documents": [documents],
+                "metadatas": [metadatas],
+            }
+        )
+
+    def _get_collection_and_tenant_id(self, collection_name: str) -> Tuple[str, str]:
+        """
+        Maps the traditional collection name to multi-tenant collection and tenant ID.
+
+        Returns:
+            tuple: (collection_name, tenant_id)
+        """
+        # Check for user memory collections
+        tenant_id = collection_name
+
+        if collection_name.startswith("user-memory-"):
+            return self.MEMORY_COLLECTION, tenant_id
+
+        # Check for file collections
+        elif collection_name.startswith("file-"):
+            return self.FILE_COLLECTION, tenant_id
+
+        # Check for web search collections
+        elif collection_name.startswith("web-search-"):
+            return self.WEB_SEARCH_COLLECTION, tenant_id
+
+        # Handle hash-based collections (YouTube and web URLs)
+        elif len(collection_name) == 63 and all(
+            c in "0123456789abcdef" for c in collection_name
+        ):
+            return self.HASH_BASED_COLLECTION, tenant_id
+
+        else:
+            return self.KNOWLEDGE_COLLECTION, tenant_id
+
+    def _extract_error_message(self, exception):
+        """
+        Extract error message from either HTTP or gRPC exceptions
+
+        Returns:
+            tuple: (status_code, error_message)
+        """
+        # Check if it's an HTTP exception
+        if isinstance(exception, UnexpectedResponse):
+            try:
+                error_data = exception.structured()
+                error_msg = error_data.get("status", {}).get("error", "")
+                return exception.status_code, error_msg
+            except Exception as inner_e:
+                log.error(f"Failed to parse HTTP error: {inner_e}")
+                return exception.status_code, str(exception)
+
+        # Check if it's a gRPC exception
+        elif isinstance(exception, grpc.RpcError):
+            # Extract status code from gRPC error
+            status_code = None
+            if hasattr(exception, "code") and callable(exception.code):
+                status_code = exception.code().value[0]
+
+            # Extract error message
+            error_msg = str(exception)
+            if "details =" in error_msg:
+                # Parse the details line which contains the actual error message
+                try:
+                    details_line = [
+                        line.strip()
+                        for line in error_msg.split("\n")
+                        if "details =" in line
+                    ][0]
+                    error_msg = details_line.split("details =")[1].strip(' "')
+                except (IndexError, AttributeError):
+                    # Fall back to full message if parsing fails
+                    pass
+
+            return status_code, error_msg
+
+        # For any other type of exception
+        return None, str(exception)
+
+    def _is_collection_not_found_error(self, exception):
+        """
+        Check if the exception is due to collection not found, supporting both HTTP and gRPC
+        """
+        status_code, error_msg = self._extract_error_message(exception)
+
+        # HTTP error (404)
+        if (
+            status_code == 404
+            and "Collection" in error_msg
+            and "doesn't exist" in error_msg
+        ):
+            return True
+
+        # gRPC error (NOT_FOUND status)
+        if (
+            isinstance(exception, grpc.RpcError)
+            and exception.code() == grpc.StatusCode.NOT_FOUND
+        ):
+            return True
+
+        return False
+
+    def _is_dimension_mismatch_error(self, exception):
+        """
+        Check if the exception is due to dimension mismatch, supporting both HTTP and gRPC
+        """
+        status_code, error_msg = self._extract_error_message(exception)
+
+        # Common patterns in both HTTP and gRPC
+        return (
+            "Vector dimension error" in error_msg
+            or "dimensions mismatch" in error_msg
+            or "invalid vector size" in error_msg
+        )
+
+    def _create_multi_tenant_collection_if_not_exists(
+        self, mt_collection_name: str, dimension: int = 384
+    ):
+        """
+        Creates a collection with multi-tenancy configuration if it doesn't exist.
+        Default dimension is set to 384 which corresponds to 'sentence-transformers/all-MiniLM-L6-v2'.
+        When creating collections dynamically (insert/upsert), the actual vector dimensions will be used.
+        """
+        try:
+            # Try to create the collection directly - will fail if it already exists
+            self.client.create_collection(
+                collection_name=mt_collection_name,
+                vectors_config=models.VectorParams(
+                    size=dimension,
+                    distance=models.Distance.COSINE,
+                    on_disk=self.QDRANT_ON_DISK,
+                ),
+                hnsw_config=models.HnswConfigDiff(
+                    payload_m=16,  # Enable per-tenant indexing
+                    m=0,
+                    on_disk=self.QDRANT_ON_DISK,
+                ),
+            )
+
+            # Create tenant ID payload index
+            self.client.create_payload_index(
+                collection_name=mt_collection_name,
+                field_name="tenant_id",
+                field_schema=models.KeywordIndexParams(
+                    type=models.KeywordIndexType.KEYWORD,
+                    is_tenant=True,
+                    on_disk=self.QDRANT_ON_DISK,
+                ),
+                wait=True,
+            )
+
+            log.info(
+                f"Multi-tenant collection {mt_collection_name} created with dimension {dimension}!"
+            )
+        except (UnexpectedResponse, grpc.RpcError) as e:
+            # Check for the specific error indicating collection already exists
+            status_code, error_msg = self._extract_error_message(e)
+
+            # HTTP status code 409 or gRPC ALREADY_EXISTS
+            if (isinstance(e, UnexpectedResponse) and status_code == 409) or (
+                isinstance(e, grpc.RpcError)
+                and e.code() == grpc.StatusCode.ALREADY_EXISTS
+            ):
+                if "already exists" in error_msg:
+                    log.debug(f"Collection {mt_collection_name} already exists")
+                    return
+            # If it's not an already exists error, re-raise
+            raise e
+        except Exception as e:
+            raise e
+
+    def _create_points(self, items: list[VectorItem], tenant_id: str):
+        """
+        Create point structs from vector items with tenant ID.
+        """
+        return [
+            PointStruct(
+                id=item["id"],
+                vector=item["vector"],
+                payload={
+                    "text": item["text"],
+                    "metadata": item["metadata"],
+                    "tenant_id": tenant_id,
+                },
+            )
+            for item in items
+        ]
+
+    def has_collection(self, collection_name: str) -> bool:
+        """
+        Check if a logical collection exists by checking for any points with the tenant ID.
+        """
+        if not self.client:
+            return False
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        # Create tenant filter
+        tenant_filter = models.FieldCondition(
+            key="tenant_id", match=models.MatchValue(value=tenant_id)
+        )
+
+        try:
+            # Try directly querying - most of the time collection should exist
+            response = self.client.query_points(
+                collection_name=mt_collection,
+                query_filter=models.Filter(must=[tenant_filter]),
+                limit=1,
+            )
+
+            # Collection exists with this tenant ID if there are points
+            return len(response.points) > 0
+        except (UnexpectedResponse, grpc.RpcError) as e:
+            if self._is_collection_not_found_error(e):
+                log.debug(f"Collection {mt_collection} doesn't exist")
+                return False
+            else:
+                # For other API errors, log and return False
+                _, error_msg = self._extract_error_message(e)
+                log.warning(f"Unexpected Qdrant error: {error_msg}")
+                return False
+        except Exception as e:
+            # For any other errors, log and return False
+            log.debug(f"Error checking collection {mt_collection}: {e}")
+            return False
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[list[str]] = None,
+        filter: Optional[dict] = None,
+    ):
+        """
+        Delete vectors by ID or filter from a collection with tenant isolation.
+        """
+        if not self.client:
+            return None
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        # Create tenant filter
+        tenant_filter = models.FieldCondition(
+            key="tenant_id", match=models.MatchValue(value=tenant_id)
+        )
+
+        must_conditions = [tenant_filter]
+        should_conditions = []
+
+        if ids:
+            for id_value in ids:
+                should_conditions.append(
+                    models.FieldCondition(
+                        key="metadata.id",
+                        match=models.MatchValue(value=id_value),
+                    ),
+                )
+        elif filter:
+            for key, value in filter.items():
+                must_conditions.append(
+                    models.FieldCondition(
+                        key=f"metadata.{key}",
+                        match=models.MatchValue(value=value),
+                    ),
+                )
+
+        try:
+            # Try to delete directly - most of the time collection should exist
+            update_result = self.client.delete(
+                collection_name=mt_collection,
+                points_selector=models.FilterSelector(
+                    filter=models.Filter(must=must_conditions, should=should_conditions)
+                ),
+            )
+
+            return update_result
+        except (UnexpectedResponse, grpc.RpcError) as e:
+            if self._is_collection_not_found_error(e):
+                log.debug(
+                    f"Collection {mt_collection} doesn't exist, nothing to delete"
+                )
+                return None
+            else:
+                # For other API errors, log and re-raise
+                _, error_msg = self._extract_error_message(e)
+                log.warning(f"Unexpected Qdrant error: {error_msg}")
+                raise
+        except Exception as e:
+            # For non-Qdrant exceptions, re-raise
+            raise
+
+    def search(
+        self, collection_name: str, vectors: list[list[float | int]], limit: int
+    ) -> Optional[SearchResult]:
+        """
+        Search for the nearest neighbor items based on the vectors with tenant isolation.
+        """
+        if not self.client:
+            return None
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        # Get the vector dimension from the query vector
+        dimension = len(vectors[0]) if vectors and len(vectors) > 0 else None
+
+        try:
+            # Try the search operation directly - most of the time collection should exist
+
+            # Create tenant filter
+            tenant_filter = models.FieldCondition(
+                key="tenant_id", match=models.MatchValue(value=tenant_id)
+            )
+
+            # Ensure vector dimensions match the collection
+            collection_dim = self.client.get_collection(
+                mt_collection
+            ).config.params.vectors.size
+
+            if collection_dim != dimension:
+                if collection_dim < dimension:
+                    vectors = [vector[:collection_dim] for vector in vectors]
+                else:
+                    vectors = [
+                        vector + [0] * (collection_dim - dimension)
+                        for vector in vectors
+                    ]
+
+            # Search with tenant filter
+            prefetch_query = models.Prefetch(
+                filter=models.Filter(must=[tenant_filter]),
+                limit=NO_LIMIT,
+            )
+            query_response = self.client.query_points(
+                collection_name=mt_collection,
+                query=vectors[0],
+                prefetch=prefetch_query,
+                limit=limit,
+            )
+
+            get_result = self._result_to_get_result(query_response.points)
+            return SearchResult(
+                ids=get_result.ids,
+                documents=get_result.documents,
+                metadatas=get_result.metadatas,
+                # qdrant distance is [-1, 1], normalize to [0, 1]
+                distances=[
+                    [(point.score + 1.0) / 2.0 for point in query_response.points]
+                ],
+            )
+        except (UnexpectedResponse, grpc.RpcError) as e:
+            if self._is_collection_not_found_error(e):
+                log.debug(
+                    f"Collection {mt_collection} doesn't exist, search returns None"
+                )
+                return None
+            else:
+                # For other API errors, log and re-raise
+                _, error_msg = self._extract_error_message(e)
+                log.warning(f"Unexpected Qdrant error during search: {error_msg}")
+                raise
+        except Exception as e:
+            # For non-Qdrant exceptions, log and return None
+            log.exception(f"Error searching collection '{collection_name}': {e}")
+            return None
+
+    def query(self, collection_name: str, filter: dict, limit: Optional[int] = None):
+        """
+        Query points with filters and tenant isolation.
+        """
+        if not self.client:
+            return None
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        # Set default limit if not provided
+        if limit is None:
+            limit = NO_LIMIT
+
+        # Create tenant filter
+        tenant_filter = models.FieldCondition(
+            key="tenant_id", match=models.MatchValue(value=tenant_id)
+        )
+
+        # Create metadata filters
+        field_conditions = []
+        for key, value in filter.items():
+            field_conditions.append(
+                models.FieldCondition(
+                    key=f"metadata.{key}", match=models.MatchValue(value=value)
+                )
+            )
+
+        # Combine tenant filter with metadata filters
+        combined_filter = models.Filter(must=[tenant_filter, *field_conditions])
+
+        try:
+            # Try the query directly - most of the time collection should exist
+            points = self.client.query_points(
+                collection_name=mt_collection,
+                query_filter=combined_filter,
+                limit=limit,
+            )
+
+            return self._result_to_get_result(points.points)
+        except (UnexpectedResponse, grpc.RpcError) as e:
+            if self._is_collection_not_found_error(e):
+                log.debug(
+                    f"Collection {mt_collection} doesn't exist, query returns None"
+                )
+                return None
+            else:
+                # For other API errors, log and re-raise
+                _, error_msg = self._extract_error_message(e)
+                log.warning(f"Unexpected Qdrant error during query: {error_msg}")
+                raise
+        except Exception as e:
+            # For non-Qdrant exceptions, log and re-raise
+            log.exception(f"Error querying collection '{collection_name}': {e}")
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        """
+        Get all items in a collection with tenant isolation.
+        """
+        if not self.client:
+            return None
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        # Create tenant filter
+        tenant_filter = models.FieldCondition(
+            key="tenant_id", match=models.MatchValue(value=tenant_id)
+        )
+
+        try:
+            # Try to get points directly - most of the time collection should exist
+            points = self.client.query_points(
+                collection_name=mt_collection,
+                query_filter=models.Filter(must=[tenant_filter]),
+                limit=NO_LIMIT,
+            )
+
+            return self._result_to_get_result(points.points)
+        except (UnexpectedResponse, grpc.RpcError) as e:
+            if self._is_collection_not_found_error(e):
+                log.debug(f"Collection {mt_collection} doesn't exist, get returns None")
+                return None
+            else:
+                # For other API errors, log and re-raise
+                _, error_msg = self._extract_error_message(e)
+                log.warning(f"Unexpected Qdrant error during get: {error_msg}")
+                raise
+        except Exception as e:
+            # For non-Qdrant exceptions, log and return None
+            log.exception(f"Error getting collection '{collection_name}': {e}")
+            return None
+
+    def _handle_operation_with_error_retry(
+        self, operation_name, mt_collection, points, dimension
+    ):
+        """
+        Private helper to handle common error cases for insert and upsert operations.
+
+        Args:
+            operation_name: 'insert' or 'upsert'
+            mt_collection: The multi-tenant collection name
+            points: The vector points to insert/upsert
+            dimension: The dimension of the vectors
+
+        Returns:
+            The operation result (for upsert) or None (for insert)
+        """
+        try:
+            if operation_name == "insert":
+                self.client.upload_points(mt_collection, points)
+                return None
+            else:  # upsert
+                return self.client.upsert(mt_collection, points)
+        except (UnexpectedResponse, grpc.RpcError) as e:
+            # Handle collection not found
+            if self._is_collection_not_found_error(e):
+                log.info(
+                    f"Collection {mt_collection} doesn't exist. Creating it with dimension {dimension}."
+                )
+                # Create collection with correct dimensions from our vectors
+                self._create_multi_tenant_collection_if_not_exists(
+                    mt_collection_name=mt_collection, dimension=dimension
+                )
+                # Try operation again - no need for dimension adjustment since we just created with correct dimensions
+                if operation_name == "insert":
+                    self.client.upload_points(mt_collection, points)
+                    return None
+                else:  # upsert
+                    return self.client.upsert(mt_collection, points)
+
+            # Handle dimension mismatch
+            elif self._is_dimension_mismatch_error(e):
+                # For dimension errors, the collection must exist, so get its configuration
+                mt_collection_info = self.client.get_collection(mt_collection)
+                existing_size = mt_collection_info.config.params.vectors.size
+
+                log.info(
+                    f"Dimension mismatch: Collection {mt_collection} expects {existing_size}, got {dimension}"
+                )
+
+                if existing_size < dimension:
+                    # Truncate vectors to fit
+                    log.info(
+                        f"Truncating vectors from {dimension} to {existing_size} dimensions"
+                    )
+                    points = [
+                        PointStruct(
+                            id=point.id,
+                            vector=point.vector[:existing_size],
+                            payload=point.payload,
+                        )
+                        for point in points
+                    ]
+                elif existing_size > dimension:
+                    # Pad vectors with zeros
+                    log.info(
+                        f"Padding vectors from {dimension} to {existing_size} dimensions with zeros"
+                    )
+                    points = [
+                        PointStruct(
+                            id=point.id,
+                            vector=point.vector
+                            + [0] * (existing_size - len(point.vector)),
+                            payload=point.payload,
+                        )
+                        for point in points
+                    ]
+                # Try operation again with adjusted dimensions
+                if operation_name == "insert":
+                    self.client.upload_points(mt_collection, points)
+                    return None
+                else:  # upsert
+                    return self.client.upsert(mt_collection, points)
+            else:
+                # Not a known error we can handle, log and re-raise
+                _, error_msg = self._extract_error_message(e)
+                log.warning(f"Unhandled Qdrant error: {error_msg}")
+                raise
+        except Exception as e:
+            # For non-Qdrant exceptions, re-raise
+            raise
+
+    def insert(self, collection_name: str, items: list[VectorItem]):
+        """
+        Insert items with tenant ID.
+        """
+        if not self.client or not items:
+            return None
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        # Get dimensions from the actual vectors
+        dimension = len(items[0]["vector"]) if items else None
+
+        # Create points with tenant ID
+        points = self._create_points(items, tenant_id)
+
+        # Handle the operation with error retry
+        return self._handle_operation_with_error_retry(
+            "insert", mt_collection, points, dimension
+        )
+
+    def upsert(self, collection_name: str, items: list[VectorItem]):
+        """
+        Upsert items with tenant ID.
+        """
+        if not self.client or not items:
+            return None
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        # Get dimensions from the actual vectors
+        dimension = len(items[0]["vector"]) if items else None
+
+        # Create points with tenant ID
+        points = self._create_points(items, tenant_id)
+
+        # Handle the operation with error retry
+        return self._handle_operation_with_error_retry(
+            "upsert", mt_collection, points, dimension
+        )
+
+    def reset(self):
+        """
+        Reset the database by deleting all collections.
+        """
+        if not self.client:
+            return None
+
+        collection_names = self.client.get_collections().collections
+        for collection_name in collection_names:
+            if collection_name.name.startswith(self.collection_prefix):
+                self.client.delete_collection(collection_name=collection_name.name)
+
+    def delete_collection(self, collection_name: str):
+        """
+        Delete a collection.
+        """
+        if not self.client:
+            return None
+
+        # Map to multi-tenant collection and tenant ID
+        mt_collection, tenant_id = self._get_collection_and_tenant_id(collection_name)
+
+        tenant_filter = models.FieldCondition(
+            key="tenant_id", match=models.MatchValue(value=tenant_id)
+        )
+
+        field_conditions = [tenant_filter]
+
+        update_result = self.client.delete(
+            collection_name=mt_collection,
+            points_selector=models.FilterSelector(
+                filter=models.Filter(must=field_conditions)
+            ),
+        )
+
+        if self.client.get_collection(mt_collection).points_count == 0:
+            self.client.delete_collection(mt_collection)
+
+        return update_result
@@ -0,0 +1,55 @@
+from open_webui.retrieval.vector.main import VectorDBBase
+from open_webui.retrieval.vector.type import VectorType
+from open_webui.config import VECTOR_DB, ENABLE_QDRANT_MULTITENANCY_MODE
+
+
+class Vector:
+
+    @staticmethod
+    def get_vector(vector_type: str) -> VectorDBBase:
+        """
+        get vector db instance by vector type
+        """
+        match vector_type:
+            case VectorType.MILVUS:
+                from open_webui.retrieval.vector.dbs.milvus import MilvusClient
+
+                return MilvusClient()
+            case VectorType.QDRANT:
+                if ENABLE_QDRANT_MULTITENANCY_MODE:
+                    from open_webui.retrieval.vector.dbs.qdrant_multitenancy import (
+                        QdrantClient,
+                    )
+
+                    return QdrantClient()
+                else:
+                    from open_webui.retrieval.vector.dbs.qdrant import QdrantClient
+
+                    return QdrantClient()
+            case VectorType.PINECONE:
+                from open_webui.retrieval.vector.dbs.pinecone import PineconeClient
+
+                return PineconeClient()
+            case VectorType.OPENSEARCH:
+                from open_webui.retrieval.vector.dbs.opensearch import OpenSearchClient
+
+                return OpenSearchClient()
+            case VectorType.PGVECTOR:
+                from open_webui.retrieval.vector.dbs.pgvector import PgvectorClient
+
+                return PgvectorClient()
+            case VectorType.ELASTICSEARCH:
+                from open_webui.retrieval.vector.dbs.elasticsearch import (
+                    ElasticsearchClient,
+                )
+
+                return ElasticsearchClient()
+            case VectorType.CHROMA:
+                from open_webui.retrieval.vector.dbs.chroma import ChromaClient
+
+                return ChromaClient()
+            case _:
+                raise ValueError(f"Unsupported vector type: {vector_type}")
+
+
+VECTOR_DB_CLIENT = Vector.get_vector(VECTOR_DB)
@@ -1,5 +1,6 @@
 from pydantic import BaseModel
-from typing import Optional, List, Any
+from abc import ABC, abstractmethod
+from typing import Any, Dict, List, Optional, Union


 class VectorItem(BaseModel):
@@ -17,3 +18,69 @@ class GetResult(BaseModel):

 class SearchResult(GetResult):
    distances: Optional[List[List[float | int]]]
+
+
+class VectorDBBase(ABC):
+    """
+    Abstract base class for all vector database backends.
+
+    Implementations of this class provide methods for collection management,
+    vector insertion, deletion, similarity search, and metadata filtering.
+
+    Any custom vector database integration must inherit from this class and
+    implement all abstract methods.
+    """
+
+    @abstractmethod
+    def has_collection(self, collection_name: str) -> bool:
+        """Check if the collection exists in the vector DB."""
+        pass
+
+    @abstractmethod
+    def delete_collection(self, collection_name: str) -> None:
+        """Delete a collection from the vector DB."""
+        pass
+
+    @abstractmethod
+    def insert(self, collection_name: str, items: List[VectorItem]) -> None:
+        """Insert a list of vector items into a collection."""
+        pass
+
+    @abstractmethod
+    def upsert(self, collection_name: str, items: List[VectorItem]) -> None:
+        """Insert or update vector items in a collection."""
+        pass
+
+    @abstractmethod
+    def search(
+        self, collection_name: str, vectors: List[List[Union[float, int]]], limit: int
+    ) -> Optional[SearchResult]:
+        """Search for similar vectors in a collection."""
+        pass
+
+    @abstractmethod
+    def query(
+        self, collection_name: str, filter: Dict, limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        """Query vectors from a collection using metadata filter."""
+        pass
+
+    @abstractmethod
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        """Retrieve all vectors from a collection."""
+        pass
+
+    @abstractmethod
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[List[str]] = None,
+        filter: Optional[Dict] = None,
+    ) -> None:
+        """Delete vectors by ID or filter from a collection."""
+        pass
+
+    @abstractmethod
+    def reset(self) -> None:
+        """Reset the vector database by removing all collections or those matching a condition."""
+        pass
@@ -0,0 +1,11 @@
+from enum import StrEnum
+
+
+class VectorType(StrEnum):
+    MILVUS = "milvus"
+    QDRANT = "qdrant"
+    CHROMA = "chroma"
+    PINECONE = "pinecone"
+    ELASTICSEARCH = "elasticsearch"
+    OPENSEARCH = "opensearch"
+    PGVECTOR = "pgvector"
@@ -0,0 +1,47 @@
+import logging
+from typing import Optional, List
+
+import requests
+from open_webui.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def search_external(
+    external_url: str,
+    external_api_key: str,
+    query: str,
+    count: int,
+    filter_list: Optional[List[str]] = None,
+) -> List[SearchResult]:
+    try:
+        response = requests.post(
+            external_url,
+            headers={
+                "User-Agent": "Open WebUI (https://github.com/open-webui/open-webui) RAG Bot",
+                "Authorization": f"Bearer {external_api_key}",
+            },
+            json={
+                "query": query,
+                "count": count,
+            },
+        )
+        response.raise_for_status()
+        results = response.json()
+        if filter_list:
+            results = get_filtered_results(results, filter_list)
+        results = [
+            SearchResult(
+                link=result.get("link"),
+                title=result.get("title"),
+                snippet=result.get("snippet"),
+            )
+            for result in results[:count]
+        ]
+        log.info(f"External search results: {results}")
+        return results
+    except Exception as e:
+        log.error(f"Error in External search: {e}")
+        return []
@@ -0,0 +1,49 @@
+import logging
+from typing import Optional, List
+from urllib.parse import urljoin
+
+import requests
+from open_webui.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def search_firecrawl(
+    firecrawl_url: str,
+    firecrawl_api_key: str,
+    query: str,
+    count: int,
+    filter_list: Optional[List[str]] = None,
+) -> List[SearchResult]:
+    try:
+        firecrawl_search_url = urljoin(firecrawl_url, "/v1/search")
+        response = requests.post(
+            firecrawl_search_url,
+            headers={
+                "User-Agent": "Open WebUI (https://github.com/open-webui/open-webui) RAG Bot",
+                "Authorization": f"Bearer {firecrawl_api_key}",
+            },
+            json={
+                "query": query,
+                "limit": count,
+            },
+        )
+        response.raise_for_status()
+        results = response.json().get("data", [])
+        if filter_list:
+            results = get_filtered_results(results, filter_list)
+        results = [
+            SearchResult(
+                link=result.get("url"),
+                title=result.get("title"),
+                snippet=result.get("description"),
+            )
+            for result in results[:count]
+        ]
+        log.info(f"External search results: {results}")
+        return results
+    except Exception as e:
+        log.error(f"Error in External search: {e}")
+        return []
@@ -2,7 +2,7 @@ import logging
 from typing import Optional

 import requests
-from open_webui.retrieval.web.main import SearchResult
+from open_webui.retrieval.web.main import SearchResult, get_filtered_results
 from open_webui.env import SRC_LOG_LEVELS

 log = logging.getLogger(__name__)
@@ -21,18 +21,25 @@ def search_tavily(
    Args:
        api_key (str): A Tavily Search API key
        query (str): The query to search for
+        count (int): The maximum number of results to return

    Returns:
        list[SearchResult]: A list of search results
    """
    url = "https://api.tavily.com/search"
-    data = {"query": query, "api_key": api_key}
-    response = requests.post(url, json=data)
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {api_key}",
+    }
+    data = {"query": query, "max_results": count}
+    response = requests.post(url, headers=headers, json=data)
    response.raise_for_status()

    json_response = response.json()

-    raw_search_results = json_response.get("results", [])
+    results = json_response.get("results", [])
+    if filter_list:
+        results = get_filtered_results(results, filter_list)

    return [
        SearchResult(
@@ -40,5 +47,5 @@ def search_tavily(
            title=result.get("title", ""),
            snippet=result.get("content"),
        )
-        for result in raw_search_results[:count]
+        for result in results
    ]
@@ -25,6 +25,7 @@ from langchain_community.document_loaders.firecrawl import FireCrawlLoader
 from langchain_community.document_loaders.base import BaseLoader
 from langchain_core.documents import Document
 from open_webui.retrieval.loaders.tavily import TavilyLoader
+from open_webui.retrieval.loaders.external_web import ExternalWebLoader
 from open_webui.constants import ERROR_MESSAGES
 from open_webui.config import (
    ENABLE_RAG_LOCAL_WEB_FETCH,
@@ -35,8 +36,10 @@ from open_webui.config import (
    FIRECRAWL_API_KEY,
    TAVILY_API_KEY,
    TAVILY_EXTRACT_DEPTH,
+    EXTERNAL_WEB_LOADER_URL,
+    EXTERNAL_WEB_LOADER_API_KEY,
 )
-from open_webui.env import SRC_LOG_LEVELS
+from open_webui.env import SRC_LOG_LEVELS, AIOHTTP_CLIENT_SESSION_SSL

 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
@@ -167,7 +170,7 @@ class SafeFireCrawlLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
        continue_on_failure: bool = True,
        api_key: Optional[str] = None,
        api_url: Optional[str] = None,
-        mode: Literal["crawl", "scrape", "map"] = "crawl",
+        mode: Literal["crawl", "scrape", "map"] = "scrape",
        proxy: Optional[Dict[str, str]] = None,
        params: Optional[Dict] = None,
    ):
@@ -225,7 +228,10 @@ class SafeFireCrawlLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
                    mode=self.mode,
                    params=self.params,
                )
-                yield from loader.lazy_load()
+                for document in loader.lazy_load():
+                    if not document.metadata.get("source"):
+                        document.metadata["source"] = document.metadata.get("sourceURL")
+                    yield document
            except Exception as e:
                if self.continue_on_failure:
                    log.exception(f"Error loading {url}: {e}")
@@ -245,6 +251,8 @@ class SafeFireCrawlLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
                    params=self.params,
                )
                async for document in loader.alazy_load():
+                    if not document.metadata.get("source"):
+                        document.metadata["source"] = document.metadata.get("sourceURL")
                    yield document
            except Exception as e:
                if self.continue_on_failure:
@@ -507,7 +515,9 @@ class SafeWebBaseLoader(WebBaseLoader):
                        kwargs["ssl"] = False

                    async with session.get(
-                        url, **(self.requests_kwargs | kwargs)
+                        url,
+                        **(self.requests_kwargs | kwargs),
+                        ssl=AIOHTTP_CLIENT_SESSION_SSL,
                    ) as response:
                        if self.raise_for_status:
                            response.raise_for_status()
@@ -619,6 +629,11 @@ def get_web_loader(
        web_loader_args["api_key"] = TAVILY_API_KEY.value
        web_loader_args["extract_depth"] = TAVILY_EXTRACT_DEPTH.value

+    if WEB_LOADER_ENGINE.value == "external":
+        WebLoaderClass = ExternalWebLoader
+        web_loader_args["external_url"] = EXTERNAL_WEB_LOADER_URL.value
+        web_loader_args["external_api_key"] = EXTERNAL_WEB_LOADER_API_KEY.value
+
    if WebLoaderClass:
        web_loader = WebLoaderClass(**web_loader_args)

@@ -0,0 +1,87 @@
+import logging
+from typing import Optional
+
+import requests
+from requests.auth import HTTPDigestAuth
+from open_webui.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def search_yacy(
+    query_url: str,
+    username: Optional[str],
+    password: Optional[str],
+    query: str,
+    count: int,
+    filter_list: Optional[list[str]] = None,
+) -> list[SearchResult]:
+    """
+    Search a Yacy instance for a given query and return the results as a list of SearchResult objects.
+
+    The function accepts username and password for authenticating to Yacy.
+
+    Args:
+        query_url (str): The base URL of the Yacy server.
+        username (str): Optional YaCy username.
+        password (str): Optional YaCy password.
+        query (str): The search term or question to find in the Yacy database.
+        count (int): The maximum number of results to retrieve from the search.
+
+    Returns:
+        list[SearchResult]: A list of SearchResults sorted by relevance score in descending order.
+
+    Raise:
+        requests.exceptions.RequestException: If a request error occurs during the search process.
+    """
+
+    # Use authentication if either username or password is set
+    yacy_auth = None
+    if username or password:
+        yacy_auth = HTTPDigestAuth(username, password)
+
+    params = {
+        "query": query,
+        "contentdom": "text",
+        "resource": "global",
+        "maximumRecords": count,
+        "nav": "none",
+    }
+
+    # Check if provided a json API URL
+    if not query_url.endswith("yacysearch.json"):
+        # Strip all query parameters from the URL
+        query_url = query_url.rstrip("/") + "/yacysearch.json"
+
+    log.debug(f"searching {query_url}")
+
+    response = requests.get(
+        query_url,
+        auth=yacy_auth,
+        headers={
+            "User-Agent": "Open WebUI (https://github.com/open-webui/open-webui) RAG Bot",
+            "Accept": "text/html",
+            "Accept-Encoding": "gzip, deflate",
+            "Accept-Language": "en-US,en;q=0.5",
+            "Connection": "keep-alive",
+        },
+        params=params,
+    )
+
+    response.raise_for_status()  # Raise an exception for HTTP errors.
+
+    json_response = response.json()
+    results = json_response.get("channels", [{}])[0].get("items", [])
+    sorted_results = sorted(results, key=lambda x: x.get("ranking", 0), reverse=True)
+    if filter_list:
+        sorted_results = get_filtered_results(sorted_results, filter_list)
+    return [
+        SearchResult(
+            link=result["link"],
+            title=result.get("title"),
+            snippet=result.get("description"),
+        )
+        for result in sorted_results[:count]
+    ]
@@ -7,6 +7,7 @@ from functools import lru_cache
 from pathlib import Path
 from pydub import AudioSegment
 from pydub.silence import split_on_silence
+from concurrent.futures import ThreadPoolExecutor

 import aiohttp
 import aiofiles
@@ -33,10 +34,12 @@ from open_webui.config import (
    WHISPER_MODEL_AUTO_UPDATE,
    WHISPER_MODEL_DIR,
    CACHE_DIR,
+    WHISPER_LANGUAGE,
 )

 from open_webui.constants import ERROR_MESSAGES
 from open_webui.env import (
+    AIOHTTP_CLIENT_SESSION_SSL,
    AIOHTTP_CLIENT_TIMEOUT,
    ENV,
    SRC_LOG_LEVELS,
@@ -48,7 +51,7 @@ from open_webui.env import (
 router = APIRouter()

 # Constants
-MAX_FILE_SIZE_MB = 25
+MAX_FILE_SIZE_MB = 20
 MAX_FILE_SIZE = MAX_FILE_SIZE_MB * 1024 * 1024  # Convert MB to bytes
 AZURE_MAX_FILE_SIZE_MB = 200
 AZURE_MAX_FILE_SIZE = AZURE_MAX_FILE_SIZE_MB * 1024 * 1024  # Convert MB to bytes
@@ -70,31 +73,50 @@ from pydub import AudioSegment
 from pydub.utils import mediainfo


-def get_audio_format(file_path):
-    """Check if the given file needs to be converted to a different format."""
+def is_audio_conversion_required(file_path):
+    """
+    Check if the given audio file needs conversion to mp3.
+    """
+    SUPPORTED_FORMATS = {"flac", "m4a", "mp3", "mp4", "mpeg", "wav", "webm"}
+
    if not os.path.isfile(file_path):
        log.error(f"File not found: {file_path}")
        return False

-    info = mediainfo(file_path)
-    if (
-        info.get("codec_name") == "aac"
-        and info.get("codec_type") == "audio"
-        and info.get("codec_tag_string") == "mp4a"
-    ):
-        return "mp4"
-    elif info.get("format_name") == "ogg":
-        return "ogg"
-    elif info.get("format_name") == "matroska,webm":
-        return "webm"
-    return None
+    try:
+        info = mediainfo(file_path)
+        codec_name = info.get("codec_name", "").lower()
+        codec_type = info.get("codec_type", "").lower()
+        codec_tag_string = info.get("codec_tag_string", "").lower()
+
+        if codec_name == "aac" and codec_type == "audio" and codec_tag_string == "mp4a":
+            # File is AAC/mp4a audio, recommend mp3 conversion
+            return True
+
+        # If the codec name or file extension is in the supported formats
+        if (
+            codec_name in SUPPORTED_FORMATS
+            or os.path.splitext(file_path)[1][1:].lower() in SUPPORTED_FORMATS
+        ):
+            return False  # Already supported
+
+        return True
+    except Exception as e:
+        log.error(f"Error getting audio format: {e}")
+        return False


-def convert_audio_to_wav(file_path, output_path, conversion_type):
-    """Convert MP4/OGG audio file to WAV format."""
-    audio = AudioSegment.from_file(file_path, format=conversion_type)
-    audio.export(output_path, format="wav")
-    log.info(f"Converted {file_path} to {output_path}")
+def convert_audio_to_mp3(file_path):
+    """Convert audio file to mp3 format."""
+    try:
+        output_path = os.path.splitext(file_path)[0] + ".mp3"
+        audio = AudioSegment.from_file(file_path)
+        audio.export(output_path, format="mp3")
+        log.info(f"Converted {file_path} to {output_path}")
+        return output_path
+    except Exception as e:
+        log.error(f"Error converting audio file: {e}")
+        return None


 def set_faster_whisper_model(model: str, auto_update: bool = False):
@@ -137,6 +159,7 @@ class TTSConfigForm(BaseModel):
    VOICE: str
    SPLIT_ON: str
    AZURE_SPEECH_REGION: str
+    AZURE_SPEECH_BASE_URL: str
    AZURE_SPEECH_OUTPUT_FORMAT: str


@@ -150,6 +173,8 @@ class STTConfigForm(BaseModel):
    AZURE_API_KEY: str
    AZURE_REGION: str
    AZURE_LOCALES: str
+    AZURE_BASE_URL: str
+    AZURE_MAX_SPEAKERS: str


 class AudioConfigUpdateForm(BaseModel):
@@ -169,6 +194,7 @@ async def get_audio_config(request: Request, user=Depends(get_admin_user)):
            "VOICE": request.app.state.config.TTS_VOICE,
            "SPLIT_ON": request.app.state.config.TTS_SPLIT_ON,
            "AZURE_SPEECH_REGION": request.app.state.config.TTS_AZURE_SPEECH_REGION,
+            "AZURE_SPEECH_BASE_URL": request.app.state.config.TTS_AZURE_SPEECH_BASE_URL,
            "AZURE_SPEECH_OUTPUT_FORMAT": request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
        },
        "stt": {
@@ -181,6 +207,8 @@ async def get_audio_config(request: Request, user=Depends(get_admin_user)):
            "AZURE_API_KEY": request.app.state.config.AUDIO_STT_AZURE_API_KEY,
            "AZURE_REGION": request.app.state.config.AUDIO_STT_AZURE_REGION,
            "AZURE_LOCALES": request.app.state.config.AUDIO_STT_AZURE_LOCALES,
+            "AZURE_BASE_URL": request.app.state.config.AUDIO_STT_AZURE_BASE_URL,
+            "AZURE_MAX_SPEAKERS": request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS,
        },
    }

@@ -197,6 +225,9 @@ async def update_audio_config(
    request.app.state.config.TTS_VOICE = form_data.tts.VOICE
    request.app.state.config.TTS_SPLIT_ON = form_data.tts.SPLIT_ON
    request.app.state.config.TTS_AZURE_SPEECH_REGION = form_data.tts.AZURE_SPEECH_REGION
+    request.app.state.config.TTS_AZURE_SPEECH_BASE_URL = (
+        form_data.tts.AZURE_SPEECH_BASE_URL
+    )
    request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT = (
        form_data.tts.AZURE_SPEECH_OUTPUT_FORMAT
    )
@@ -210,6 +241,10 @@ async def update_audio_config(
    request.app.state.config.AUDIO_STT_AZURE_API_KEY = form_data.stt.AZURE_API_KEY
    request.app.state.config.AUDIO_STT_AZURE_REGION = form_data.stt.AZURE_REGION
    request.app.state.config.AUDIO_STT_AZURE_LOCALES = form_data.stt.AZURE_LOCALES
+    request.app.state.config.AUDIO_STT_AZURE_BASE_URL = form_data.stt.AZURE_BASE_URL
+    request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS = (
+        form_data.stt.AZURE_MAX_SPEAKERS
+    )

    if request.app.state.config.STT_ENGINE == "":
        request.app.state.faster_whisper_model = set_faster_whisper_model(
@@ -226,6 +261,7 @@ async def update_audio_config(
            "VOICE": request.app.state.config.TTS_VOICE,
            "SPLIT_ON": request.app.state.config.TTS_SPLIT_ON,
            "AZURE_SPEECH_REGION": request.app.state.config.TTS_AZURE_SPEECH_REGION,
+            "AZURE_SPEECH_BASE_URL": request.app.state.config.TTS_AZURE_SPEECH_BASE_URL,
            "AZURE_SPEECH_OUTPUT_FORMAT": request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
        },
        "stt": {
@@ -238,6 +274,8 @@ async def update_audio_config(
            "AZURE_API_KEY": request.app.state.config.AUDIO_STT_AZURE_API_KEY,
            "AZURE_REGION": request.app.state.config.AUDIO_STT_AZURE_REGION,
            "AZURE_LOCALES": request.app.state.config.AUDIO_STT_AZURE_LOCALES,
+            "AZURE_BASE_URL": request.app.state.config.AUDIO_STT_AZURE_BASE_URL,
+            "AZURE_MAX_SPEAKERS": request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS,
        },
    }

@@ -305,6 +343,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
                            else {}
                        ),
                    },
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
                ) as r:
                    r.raise_for_status()

@@ -360,6 +399,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
                        "Content-Type": "application/json",
                        "xi-api-key": request.app.state.config.TTS_API_KEY,
                    },
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
                ) as r:
                    r.raise_for_status()

@@ -395,7 +435,8 @@ async def speech(request: Request, user=Depends(get_verified_user)):
            log.exception(e)
            raise HTTPException(status_code=400, detail="Invalid JSON payload")

-        region = request.app.state.config.TTS_AZURE_SPEECH_REGION
+        region = request.app.state.config.TTS_AZURE_SPEECH_REGION or "eastus"
+        base_url = request.app.state.config.TTS_AZURE_SPEECH_BASE_URL
        language = request.app.state.config.TTS_VOICE
        locale = "-".join(request.app.state.config.TTS_VOICE.split("-")[:1])
        output_format = request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT
@@ -409,13 +450,15 @@ async def speech(request: Request, user=Depends(get_verified_user)):
                timeout=timeout, trust_env=True
            ) as session:
                async with session.post(
-                    f"https://{region}.tts.speech.microsoft.com/cognitiveservices/v1",
+                    (base_url or f"https://{region}.tts.speech.microsoft.com")
+                    + "/cognitiveservices/v1",
                    headers={
                        "Ocp-Apim-Subscription-Key": request.app.state.config.TTS_API_KEY,
                        "Content-Type": "application/ssml+xml",
                        "X-Microsoft-OutputFormat": output_format,
                    },
                    data=data,
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
                ) as r:
                    r.raise_for_status()

@@ -484,8 +527,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
        return FileResponse(file_path)


-def transcribe(request: Request, file_path):
-    log.info(f"transcribe: {file_path}")
+def transcription_handler(request, file_path):
    filename = os.path.basename(file_path)
    file_dir = os.path.dirname(file_path)
    id = filename.split(".")[0]
@@ -497,7 +539,12 @@ def transcribe(request: Request, file_path):
            )

        model = request.app.state.faster_whisper_model
-        segments, info = model.transcribe(file_path, beam_size=5)
+        segments, info = model.transcribe(
+            file_path,
+            beam_size=5,
+            vad_filter=request.app.state.config.WHISPER_VAD_FILTER,
+            language=WHISPER_LANGUAGE,
+        )
        log.info(
            "Detected language '%s' with probability %f"
            % (info.language, info.language_probability)
@@ -514,16 +561,6 @@ def transcribe(request: Request, file_path):
        log.debug(data)
        return data
    elif request.app.state.config.STT_ENGINE == "openai":
-        audio_format = get_audio_format(file_path)
-        if audio_format:
-            os.rename(file_path, file_path.replace(".wav", f".{audio_format}"))
-            # Convert unsupported audio file to WAV format
-            convert_audio_to_wav(
-                file_path.replace(".wav", f".{audio_format}"),
-                file_path,
-                audio_format,
-            )
-
        r = None
        try:
            r = requests.post(
@@ -635,8 +672,10 @@ def transcribe(request: Request, file_path):
            )

        api_key = request.app.state.config.AUDIO_STT_AZURE_API_KEY
-        region = request.app.state.config.AUDIO_STT_AZURE_REGION
+        region = request.app.state.config.AUDIO_STT_AZURE_REGION or "eastus"
        locales = request.app.state.config.AUDIO_STT_AZURE_LOCALES
+        base_url = request.app.state.config.AUDIO_STT_AZURE_BASE_URL
+        max_speakers = request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS or 3

        # IF NO LOCALES, USE DEFAULTS
        if len(locales) < 2:
@@ -660,7 +699,7 @@ def transcribe(request: Request, file_path):
        if not api_key or not region:
            raise HTTPException(
                status_code=400,
-                detail="Azure API key and region are required for Azure STT",
+                detail="Azure API key is required for Azure STT",
            )

        r = None
@@ -670,13 +709,16 @@ def transcribe(request: Request, file_path):
                "definition": json.dumps(
                    {
                        "locales": locales.split(","),
-                        "diarization": {"maxSpeakers": 3, "enabled": True},
+                        "diarization": {"maxSpeakers": max_speakers, "enabled": True},
                    }
                    if locales
                    else {}
                )
            }
-            url = f"https://{region}.api.cognitive.microsoft.com/speechtotext/transcriptions:transcribe?api-version=2024-11-15"
+
+            url = (
+                base_url or f"https://{region}.api.cognitive.microsoft.com"
+            ) + "/speechtotext/transcriptions:transcribe?api-version=2024-11-15"

            # Use context manager to ensure file is properly closed
            with open(file_path, "rb") as audio_file:
@@ -735,24 +777,122 @@ def transcribe(request: Request, file_path):
            )


+def transcribe(request: Request, file_path):
+    log.info(f"transcribe: {file_path}")
+
+    if is_audio_conversion_required(file_path):
+        file_path = convert_audio_to_mp3(file_path)
+
+    try:
+        file_path = compress_audio(file_path)
+    except Exception as e:
+        log.exception(e)
+
+    # Always produce a list of chunk paths (could be one entry if small)
+    try:
+        chunk_paths = split_audio(file_path, MAX_FILE_SIZE)
+        print(f"Chunk paths: {chunk_paths}")
+    except Exception as e:
+        log.exception(e)
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.DEFAULT(e),
+        )
+
+    results = []
+    try:
+        with ThreadPoolExecutor() as executor:
+            # Submit tasks for each chunk_path
+            futures = [
+                executor.submit(transcription_handler, request, chunk_path)
+                for chunk_path in chunk_paths
+            ]
+            # Gather results as they complete
+            for future in futures:
+                try:
+                    results.append(future.result())
+                except Exception as transcribe_exc:
+                    log.exception(f"Error transcribing chunk: {transcribe_exc}")
+                    raise HTTPException(
+                        status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+                        detail="Error during transcription.",
+                    )
+    finally:
+        # Clean up only the temporary chunks, never the original file
+        for chunk_path in chunk_paths:
+            if chunk_path != file_path and os.path.isfile(chunk_path):
+                try:
+                    os.remove(chunk_path)
+                except Exception:
+                    pass
+
+    return {
+        "text": " ".join([result["text"] for result in results]),
+    }
+
+
 def compress_audio(file_path):
    if os.path.getsize(file_path) > MAX_FILE_SIZE:
+        id = os.path.splitext(os.path.basename(file_path))[
+            0
+        ]  # Handles names with multiple dots
        file_dir = os.path.dirname(file_path)
+
        audio = AudioSegment.from_file(file_path)
        audio = audio.set_frame_rate(16000).set_channels(1)  # Compress audio
-        compressed_path = f"{file_dir}/{id}_compressed.opus"
-        audio.export(compressed_path, format="opus", bitrate="32k")
-        log.debug(f"Compressed audio to {compressed_path}")

-        if (
-            os.path.getsize(compressed_path) > MAX_FILE_SIZE
-        ):  # Still larger than MAX_FILE_SIZE after compression
-            raise Exception(ERROR_MESSAGES.FILE_TOO_LARGE(size=f"{MAX_FILE_SIZE_MB}MB"))
+        compressed_path = os.path.join(file_dir, f"{id}_compressed.mp3")
+        audio.export(compressed_path, format="mp3", bitrate="32k")
+        # log.debug(f"Compressed audio to {compressed_path}")  # Uncomment if log is defined
+
        return compressed_path
    else:
        return file_path


+def split_audio(file_path, max_bytes, format="mp3", bitrate="32k"):
+    """
+    Splits audio into chunks not exceeding max_bytes.
+    Returns a list of chunk file paths. If audio fits, returns list with original path.
+    """
+    file_size = os.path.getsize(file_path)
+    if file_size <= max_bytes:
+        return [file_path]  # Nothing to split
+
+    audio = AudioSegment.from_file(file_path)
+    duration_ms = len(audio)
+    orig_size = file_size
+
+    approx_chunk_ms = max(int(duration_ms * (max_bytes / orig_size)) - 1000, 1000)
+    chunks = []
+    start = 0
+    i = 0
+
+    base, _ = os.path.splitext(file_path)
+
+    while start < duration_ms:
+        end = min(start + approx_chunk_ms, duration_ms)
+        chunk = audio[start:end]
+        chunk_path = f"{base}_chunk_{i}.{format}"
+        chunk.export(chunk_path, format=format, bitrate=bitrate)
+
+        # Reduce chunk duration if still too large
+        while os.path.getsize(chunk_path) > max_bytes and (end - start) > 5000:
+            end = start + ((end - start) // 2)
+            chunk = audio[start:end]
+            chunk.export(chunk_path, format=format, bitrate=bitrate)
+
+        if os.path.getsize(chunk_path) > max_bytes:
+            os.remove(chunk_path)
+            raise Exception("Audio chunk cannot be reduced below max file size.")
+
+        chunks.append(chunk_path)
+        start = end
+        i += 1
+
+    return chunks
+
+
@router.post("/transcriptions")
 def transcription(
    request: Request,
@@ -761,9 +901,11 @@ def transcription(
 ):
    log.info(f"file.content_type: {file.content_type}")

-    supported_filetypes = ("audio/mpeg", "audio/wav", "audio/ogg", "audio/x-m4a")
-
-    if not file.content_type.startswith(supported_filetypes):
+    SUPPORTED_CONTENT_TYPES = {"video/webm"}  # Extend if you add more video types!
+    if not (
+        file.content_type.startswith("audio/")
+        or file.content_type in SUPPORTED_CONTENT_TYPES
+    ):
        raise HTTPException(
            status_code=status.HTTP_400_BAD_REQUEST,
            detail=ERROR_MESSAGES.FILE_NOT_SUPPORTED,
@@ -784,19 +926,13 @@ def transcription(
            f.write(contents)

        try:
-            try:
-                file_path = compress_audio(file_path)
-            except Exception as e:
-                log.exception(e)
+            result = transcribe(request, file_path)

-                raise HTTPException(
-                    status_code=status.HTTP_400_BAD_REQUEST,
-                    detail=ERROR_MESSAGES.DEFAULT(e),
-                )
+            return {
+                **result,
+                "filename": os.path.basename(file_path),
+            }

-            data = transcribe(request, file_path)
-            file_path = file_path.split("/")[-1]
-            return {**data, "filename": file_path}
        except Exception as e:
            log.exception(e)

@@ -905,7 +1041,10 @@ def get_available_voices(request) -> dict:
    elif request.app.state.config.TTS_ENGINE == "azure":
        try:
            region = request.app.state.config.TTS_AZURE_SPEECH_REGION
-            url = f"https://{region}.tts.speech.microsoft.com/cognitiveservices/voices/list"
+            base_url = request.app.state.config.TTS_AZURE_SPEECH_BASE_URL
+            url = (
+                base_url or f"https://{region}.tts.speech.microsoft.com"
+            ) + "/cognitiveservices/voices/list"
            headers = {
                "Ocp-Apim-Subscription-Key": request.app.state.config.TTS_API_KEY
            }
@@ -27,27 +27,31 @@ from open_webui.env import (
    WEBUI_AUTH_TRUSTED_NAME_HEADER,
    WEBUI_AUTH_COOKIE_SAME_SITE,
    WEBUI_AUTH_COOKIE_SECURE,
+    WEBUI_AUTH_SIGNOUT_REDIRECT_URL,
    SRC_LOG_LEVELS,
 )
 from fastapi import APIRouter, Depends, HTTPException, Request, status
-from fastapi.responses import RedirectResponse, Response
+from fastapi.responses import RedirectResponse, Response, JSONResponse
 from open_webui.config import OPENID_PROVIDER_URL, ENABLE_OAUTH_SIGNUP, ENABLE_LDAP
 from pydantic import BaseModel
+
 from open_webui.utils.misc import parse_duration, validate_email_format
 from open_webui.utils.auth import (
+    decode_token,
    create_api_key,
    create_token,
    get_admin_user,
    get_verified_user,
    get_current_user,
    get_password_hash,
+    get_http_authorization_cred,
 )
 from open_webui.utils.webhook import post_webhook
 from open_webui.utils.access_control import get_permissions

 from typing import Optional, List

-from ssl import CERT_REQUIRED, PROTOCOL_TLS
+from ssl import CERT_NONE, CERT_REQUIRED, PROTOCOL_TLS

 if ENABLE_LDAP.value:
    from ldap3 import Server, Connection, NONE, Tls
@@ -72,31 +76,36 @@ class SessionUserResponse(Token, UserResponse):
 async def get_session_user(
    request: Request, response: Response, user=Depends(get_current_user)
 ):
-    expires_delta = parse_duration(request.app.state.config.JWT_EXPIRES_IN)
+
+    auth_header = request.headers.get("Authorization")
+    auth_token = get_http_authorization_cred(auth_header)
+    token = auth_token.credentials
+    data = decode_token(token)
+
    expires_at = None
-    if expires_delta:
-        expires_at = int(time.time()) + int(expires_delta.total_seconds())

-    token = create_token(
-        data={"id": user.id},
-        expires_delta=expires_delta,
-    )
+    if data:
+        expires_at = data.get("exp")

-    datetime_expires_at = (
-        datetime.datetime.fromtimestamp(expires_at, datetime.timezone.utc)
-        if expires_at
-        else None
-    )
+        if (expires_at is not None) and int(time.time()) > expires_at:
+            raise HTTPException(
+                status_code=status.HTTP_401_UNAUTHORIZED,
+                detail=ERROR_MESSAGES.INVALID_TOKEN,
+            )

-    # Set the cookie token
-    response.set_cookie(
-        key="token",
-        value=token,
-        expires=datetime_expires_at,
-        httponly=True,  # Ensures the cookie is not accessible via JavaScript
-        samesite=WEBUI_AUTH_COOKIE_SAME_SITE,
-        secure=WEBUI_AUTH_COOKIE_SECURE,
-    )
+        # Set the cookie token
+        response.set_cookie(
+            key="token",
+            value=token,
+            expires=(
+                datetime.datetime.fromtimestamp(expires_at, datetime.timezone.utc)
+                if expires_at
+                else None
+            ),
+            httponly=True,  # Ensures the cookie is not accessible via JavaScript
+            samesite=WEBUI_AUTH_COOKIE_SAME_SITE,
+            secure=WEBUI_AUTH_COOKIE_SECURE,
+        )

    user_permissions = get_permissions(
        user.id, request.app.state.config.USER_PERMISSIONS
@@ -177,6 +186,9 @@ async def ldap_auth(request: Request, response: Response, form_data: LdapForm):
    LDAP_APP_PASSWORD = request.app.state.config.LDAP_APP_PASSWORD
    LDAP_USE_TLS = request.app.state.config.LDAP_USE_TLS
    LDAP_CA_CERT_FILE = request.app.state.config.LDAP_CA_CERT_FILE
+    LDAP_VALIDATE_CERT = (
+        CERT_REQUIRED if request.app.state.config.LDAP_VALIDATE_CERT else CERT_NONE
+    )
    LDAP_CIPHERS = (
        request.app.state.config.LDAP_CIPHERS
        if request.app.state.config.LDAP_CIPHERS
@@ -188,7 +200,7 @@ async def ldap_auth(request: Request, response: Response, form_data: LdapForm):

    try:
        tls = Tls(
-            validate=CERT_REQUIRED,
+            validate=LDAP_VALIDATE_CERT,
            version=PROTOCOL_TLS,
            ca_certs_file=LDAP_CA_CERT_FILE,
            ciphers=LDAP_CIPHERS,
@@ -225,18 +237,22 @@ async def ldap_auth(request: Request, response: Response, form_data: LdapForm):
            ],
        )

-        if not search_success:
+        if not search_success or not connection_app.entries:
            raise HTTPException(400, detail="User not found in the LDAP server")

        entry = connection_app.entries[0]
        username = str(entry[f"{LDAP_ATTRIBUTE_FOR_USERNAME}"]).lower()
-        email = entry[f"{LDAP_ATTRIBUTE_FOR_MAIL}"]
+        email = entry[
+            f"{LDAP_ATTRIBUTE_FOR_MAIL}"
+        ].value  # retrieve the Attribute value
        if not email:
            raise HTTPException(400, "User does not have a valid email address.")
        elif isinstance(email, str):
            email = email.lower()
        elif isinstance(email, list):
            email = email[0].lower()
+        else:
+            email = str(email).lower()

        cn = str(entry["cn"])
        user_dn = entry.entry_dn
@@ -286,18 +302,30 @@ async def ldap_auth(request: Request, response: Response, form_data: LdapForm):
            user = Auths.authenticate_user_by_trusted_header(email)

            if user:
+                expires_delta = parse_duration(request.app.state.config.JWT_EXPIRES_IN)
+                expires_at = None
+                if expires_delta:
+                    expires_at = int(time.time()) + int(expires_delta.total_seconds())
+
                token = create_token(
                    data={"id": user.id},
-                    expires_delta=parse_duration(
-                        request.app.state.config.JWT_EXPIRES_IN
-                    ),
+                    expires_delta=expires_delta,
                )

                # Set the cookie token
                response.set_cookie(
                    key="token",
                    value=token,
+                    expires=(
+                        datetime.datetime.fromtimestamp(
+                            expires_at, datetime.timezone.utc
+                        )
+                        if expires_at
+                        else None
+                    ),
                    httponly=True,  # Ensures the cookie is not accessible via JavaScript
+                    samesite=WEBUI_AUTH_COOKIE_SAME_SITE,
+                    secure=WEBUI_AUTH_COOKIE_SECURE,
                )

                user_permissions = get_permissions(
@@ -307,6 +335,7 @@ async def ldap_auth(request: Request, response: Response, form_data: LdapForm):
                return {
                    "token": token,
                    "token_type": "Bearer",
+                    "expires_at": expires_at,
                    "id": user.id,
                    "email": user.email,
                    "name": user.name,
@@ -452,10 +481,6 @@ async def signup(request: Request, response: Response, form_data: SignupForm):
            "admin" if user_count == 0 else request.app.state.config.DEFAULT_USER_ROLE
        )

-        if user_count == 0:
-            # Disable signup after the first user is created
-            request.app.state.config.ENABLE_SIGNUP = False
-
        # The password passed to bcrypt must be 72 bytes or fewer. If it is longer, it will be truncated before hashing.
        if len(form_data.password.encode("utf-8")) > 72:
            raise HTTPException(
@@ -515,6 +540,10 @@ async def signup(request: Request, response: Response, form_data: SignupForm):
                user.id, request.app.state.config.USER_PERMISSIONS
            )

+            if user_count == 0:
+                # Disable signup after the first user is created
+                request.app.state.config.ENABLE_SIGNUP = False
+
            return {
                "token": token,
                "token_type": "Bearer",
@@ -548,9 +577,14 @@ async def signout(request: Request, response: Response):
                            logout_url = openid_data.get("end_session_endpoint")
                            if logout_url:
                                response.delete_cookie("oauth_id_token")
-                                return RedirectResponse(
+
+                                return JSONResponse(
+                                    status_code=200,
+                                    content={
+                                        "status": True,
+                                        "redirect_url": f"{logout_url}?id_token_hint={oauth_id_token}",
+                                    },
                                    headers=response.headers,
-                                    url=f"{logout_url}?id_token_hint={oauth_id_token}",
                                )
                        else:
                            raise HTTPException(
@@ -564,7 +598,19 @@ async def signout(request: Request, response: Response):
                    detail="Failed to sign out from the OpenID provider.",
                )

-    return {"status": True}
+    if WEBUI_AUTH_SIGNOUT_REDIRECT_URL:
+        return JSONResponse(
+            status_code=200,
+            content={
+                "status": True,
+                "redirect_url": WEBUI_AUTH_SIGNOUT_REDIRECT_URL,
+            },
+            headers=response.headers,
+        )
+
+    return JSONResponse(
+        status_code=200, content={"status": True}, headers=response.headers
+    )


 ############################
@@ -662,7 +708,11 @@ async def get_admin_config(request: Request, user=Depends(get_admin_user)):
        "ENABLE_COMMUNITY_SHARING": request.app.state.config.ENABLE_COMMUNITY_SHARING,
        "ENABLE_MESSAGE_RATING": request.app.state.config.ENABLE_MESSAGE_RATING,
        "ENABLE_CHANNELS": request.app.state.config.ENABLE_CHANNELS,
+        "ENABLE_NOTES": request.app.state.config.ENABLE_NOTES,
        "ENABLE_USER_WEBHOOKS": request.app.state.config.ENABLE_USER_WEBHOOKS,
+        "PENDING_USER_OVERLAY_TITLE": request.app.state.config.PENDING_USER_OVERLAY_TITLE,
+        "PENDING_USER_OVERLAY_CONTENT": request.app.state.config.PENDING_USER_OVERLAY_CONTENT,
+        "RESPONSE_WATERMARK": request.app.state.config.RESPONSE_WATERMARK,
    }


@@ -678,7 +728,11 @@ class AdminConfig(BaseModel):
    ENABLE_COMMUNITY_SHARING: bool
    ENABLE_MESSAGE_RATING: bool
    ENABLE_CHANNELS: bool
+    ENABLE_NOTES: bool
    ENABLE_USER_WEBHOOKS: bool
+    PENDING_USER_OVERLAY_TITLE: Optional[str] = None
+    PENDING_USER_OVERLAY_CONTENT: Optional[str] = None
+    RESPONSE_WATERMARK: Optional[str] = None


@router.post("/admin/config")
@@ -698,6 +752,7 @@ async def update_admin_config(
    )

    request.app.state.config.ENABLE_CHANNELS = form_data.ENABLE_CHANNELS
+    request.app.state.config.ENABLE_NOTES = form_data.ENABLE_NOTES

    if form_data.DEFAULT_USER_ROLE in ["pending", "user", "admin"]:
        request.app.state.config.DEFAULT_USER_ROLE = form_data.DEFAULT_USER_ROLE
@@ -715,6 +770,15 @@ async def update_admin_config(

    request.app.state.config.ENABLE_USER_WEBHOOKS = form_data.ENABLE_USER_WEBHOOKS

+    request.app.state.config.PENDING_USER_OVERLAY_TITLE = (
+        form_data.PENDING_USER_OVERLAY_TITLE
+    )
+    request.app.state.config.PENDING_USER_OVERLAY_CONTENT = (
+        form_data.PENDING_USER_OVERLAY_CONTENT
+    )
+
+    request.app.state.config.RESPONSE_WATERMARK = form_data.RESPONSE_WATERMARK
+
    return {
        "SHOW_ADMIN_DETAILS": request.app.state.config.SHOW_ADMIN_DETAILS,
        "WEBUI_URL": request.app.state.config.WEBUI_URL,
@@ -722,12 +786,16 @@ async def update_admin_config(
        "ENABLE_API_KEY": request.app.state.config.ENABLE_API_KEY,
        "ENABLE_API_KEY_ENDPOINT_RESTRICTIONS": request.app.state.config.ENABLE_API_KEY_ENDPOINT_RESTRICTIONS,
        "API_KEY_ALLOWED_ENDPOINTS": request.app.state.config.API_KEY_ALLOWED_ENDPOINTS,
-        "ENABLE_CHANNELS": request.app.state.config.ENABLE_CHANNELS,
        "DEFAULT_USER_ROLE": request.app.state.config.DEFAULT_USER_ROLE,
        "JWT_EXPIRES_IN": request.app.state.config.JWT_EXPIRES_IN,
        "ENABLE_COMMUNITY_SHARING": request.app.state.config.ENABLE_COMMUNITY_SHARING,
        "ENABLE_MESSAGE_RATING": request.app.state.config.ENABLE_MESSAGE_RATING,
+        "ENABLE_CHANNELS": request.app.state.config.ENABLE_CHANNELS,
+        "ENABLE_NOTES": request.app.state.config.ENABLE_NOTES,
        "ENABLE_USER_WEBHOOKS": request.app.state.config.ENABLE_USER_WEBHOOKS,
+        "PENDING_USER_OVERLAY_TITLE": request.app.state.config.PENDING_USER_OVERLAY_TITLE,
+        "PENDING_USER_OVERLAY_CONTENT": request.app.state.config.PENDING_USER_OVERLAY_CONTENT,
+        "RESPONSE_WATERMARK": request.app.state.config.RESPONSE_WATERMARK,
    }


@@ -743,6 +811,7 @@ class LdapServerConfig(BaseModel):
    search_filters: str = ""
    use_tls: bool = True
    certificate_path: Optional[str] = None
+    validate_cert: bool = True
    ciphers: Optional[str] = "ALL"


@@ -760,6 +829,7 @@ async def get_ldap_server(request: Request, user=Depends(get_admin_user)):
        "search_filters": request.app.state.config.LDAP_SEARCH_FILTERS,
        "use_tls": request.app.state.config.LDAP_USE_TLS,
        "certificate_path": request.app.state.config.LDAP_CA_CERT_FILE,
+        "validate_cert": request.app.state.config.LDAP_VALIDATE_CERT,
        "ciphers": request.app.state.config.LDAP_CIPHERS,
    }

@@ -795,6 +865,7 @@ async def update_ldap_server(
    request.app.state.config.LDAP_SEARCH_FILTERS = form_data.search_filters
    request.app.state.config.LDAP_USE_TLS = form_data.use_tls
    request.app.state.config.LDAP_CA_CERT_FILE = form_data.certificate_path
+    request.app.state.config.LDAP_VALIDATE_CERT = form_data.validate_cert
    request.app.state.config.LDAP_CIPHERS = form_data.ciphers

    return {
@@ -809,6 +880,7 @@ async def update_ldap_server(
        "search_filters": request.app.state.config.LDAP_SEARCH_FILTERS,
        "use_tls": request.app.state.config.LDAP_USE_TLS,
        "certificate_path": request.app.state.config.LDAP_CA_CERT_FILE,
+        "validate_cert": request.app.state.config.LDAP_VALIDATE_CERT,
        "ciphers": request.app.state.config.LDAP_CIPHERS,
    }

@@ -638,8 +638,17 @@ async def archive_chat_by_id(id: str, user=Depends(get_verified_user)):


@router.post("/{id}/share", response_model=Optional[ChatResponse])
-async def share_chat_by_id(id: str, user=Depends(get_verified_user)):
+async def share_chat_by_id(request: Request, id: str, user=Depends(get_verified_user)):
+    if not has_permission(
+        user.id, "chat.share", request.app.state.config.USER_PERMISSIONS
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+        )
+
    chat = Chats.get_chat_by_id_and_user_id(id, user.id)
+
    if chat:
        if chat.share_id:
            shared_chat = Chats.update_shared_chat_by_chat_id(chat.id)
@@ -56,7 +56,7 @@ async def update_config(
    }


-class FeedbackUserReponse(BaseModel):
+class UserResponse(BaseModel):
    id: str
    name: str
    email: str
@@ -68,21 +68,23 @@ class FeedbackUserReponse(BaseModel):


 class FeedbackUserResponse(FeedbackResponse):
-    user: Optional[FeedbackUserReponse] = None
+    user: Optional[UserResponse] = None


@router.get("/feedbacks/all", response_model=list[FeedbackUserResponse])
 async def get_all_feedbacks(user=Depends(get_admin_user)):
    feedbacks = Feedbacks.get_all_feedbacks()
-    return [
-        FeedbackUserResponse(
-            **feedback.model_dump(),
-            user=FeedbackUserReponse(
-                **Users.get_user_by_id(feedback.user_id).model_dump()
-            ),
+
+    feedback_list = []
+    for feedback in feedbacks:
+        user = Users.get_user_by_id(feedback.user_id)
+        feedback_list.append(
+            FeedbackUserResponse(
+                **feedback.model_dump(),
+                user=UserResponse(**user.model_dump()) if user else None,
+            )
        )
-        for feedback in feedbacks
-    ]
+    return feedback_list


@router.delete("/feedbacks/all")
@@ -94,12 +96,7 @@ async def delete_all_feedbacks(user=Depends(get_admin_user)):
@router.get("/feedbacks/all/export", response_model=list[FeedbackModel])
 async def get_all_feedbacks(user=Depends(get_admin_user)):
    feedbacks = Feedbacks.get_all_feedbacks()
-    return [
-        FeedbackModel(
-            **feedback.model_dump(), user=Users.get_user_by_id(feedback.user_id)
-        )
-        for feedback in feedbacks
-    ]
+    return feedbacks


@router.get("/feedbacks/user", response_model=list[FeedbackUserResponse])
@@ -19,6 +19,8 @@ from fastapi import (
 from fastapi.responses import FileResponse, StreamingResponse
 from open_webui.constants import ERROR_MESSAGES
 from open_webui.env import SRC_LOG_LEVELS
+
+from open_webui.models.users import Users
 from open_webui.models.files import (
    FileForm,
    FileModel,
@@ -83,19 +85,41 @@ def upload_file(
    request: Request,
    file: UploadFile = File(...),
    user=Depends(get_verified_user),
-    file_metadata: dict = {},
+    file_metadata: dict = None,
    process: bool = Query(True),
 ):
    log.info(f"file.content_type: {file.content_type}")
+
+    file_metadata = file_metadata if file_metadata else {}
    try:
        unsanitized_filename = file.filename
        filename = os.path.basename(unsanitized_filename)

+        file_extension = os.path.splitext(filename)[1]
+        if request.app.state.config.ALLOWED_FILE_EXTENSIONS:
+            request.app.state.config.ALLOWED_FILE_EXTENSIONS = [
+                ext for ext in request.app.state.config.ALLOWED_FILE_EXTENSIONS if ext
+            ]
+
+            if file_extension not in request.app.state.config.ALLOWED_FILE_EXTENSIONS:
+                raise HTTPException(
+                    status_code=status.HTTP_400_BAD_REQUEST,
+                    detail=ERROR_MESSAGES.DEFAULT(
+                        f"File type {file_extension} is not allowed"
+                    ),
+                )
+
        # replace filename with uuid
        id = str(uuid.uuid4())
        name = filename
        filename = f"{id}_{filename}"
-        contents, file_path = Storage.upload_file(file.file, filename)
+        tags = {
+            "OpenWebUI-User-Email": user.email,
+            "OpenWebUI-User-Id": user.id,
+            "OpenWebUI-User-Name": user.name,
+            "OpenWebUI-File-Id": id,
+        }
+        contents, file_path = Storage.upload_file(file.file, filename, tags)

        file_item = Files.insert_new_file(
            user.id,
@@ -115,21 +139,31 @@ def upload_file(
        )
        if process:
            try:
-                if file.content_type in [
-                    "audio/mpeg",
-                    "audio/wav",
-                    "audio/ogg",
-                    "audio/x-m4a",
-                ]:
-                    file_path = Storage.get_file(file_path)
-                    result = transcribe(request, file_path)
+                if file.content_type:
+                    if file.content_type.startswith("audio/") or file.content_type in {
+                        "video/webm"
+                    }:
+                        file_path = Storage.get_file(file_path)
+                        result = transcribe(request, file_path)

-                    process_file(
-                        request,
-                        ProcessFileForm(file_id=id, content=result.get("text", "")),
-                        user=user,
+                        process_file(
+                            request,
+                            ProcessFileForm(file_id=id, content=result.get("text", "")),
+                            user=user,
+                        )
+                    elif file.content_type not in [
+                        "image/png",
+                        "image/jpeg",
+                        "image/gif",
+                        "video/mp4",
+                        "video/ogg",
+                        "video/quicktime",
+                    ]:
+                        process_file(request, ProcessFileForm(file_id=id), user=user)
+                else:
+                    log.info(
+                        f"File type {file.content_type} is not provided, but trying to process anyway"
                    )
-                elif file.content_type not in ["image/png", "image/jpeg", "image/gif"]:
                    process_file(request, ProcessFileForm(file_id=id), user=user)

                file_item = Files.get_file_by_id(id=id)
@@ -173,7 +207,8 @@ async def list_files(user=Depends(get_verified_user), content: bool = Query(True

    if not content:
        for file in files:
-            del file.data["content"]
+            if "content" in file.data:
+                del file.data["content"]

    return files

@@ -214,7 +249,8 @@ async def search_files(

    if not content:
        for file in matching_files:
-            del file.data["content"]
+            if "content" in file.data:
+                del file.data["content"]

    return matching_files

@@ -431,6 +467,13 @@ async def get_html_file_content_by_id(id: str, user=Depends(get_verified_user)):
            detail=ERROR_MESSAGES.NOT_FOUND,
        )

+    file_user = Users.get_user_by_id(file.user_id)
+    if not file_user.role == "admin":
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
    if (
        file.user_id == user.id
        or user.role == "admin"
@@ -500,7 +500,11 @@ async def image_generations(
                    if form_data.size
                    else request.app.state.config.IMAGE_SIZE
                ),
-                "response_format": "b64_json",
+                **(
+                    {}
+                    if "gpt-image-1" in request.app.state.config.IMAGE_GENERATION_MODEL
+                    else {"response_format": "b64_json"}
+                ),
            }

            # Use asyncio.to_thread for the requests.post call
@@ -619,7 +623,7 @@ async def image_generations(
            or request.app.state.config.IMAGE_GENERATION_ENGINE == ""
        ):
            if form_data.model:
-                set_image_model(form_data.model)
+                set_image_model(request, form_data.model)

            data = {
                "prompt": form_data.prompt,
@@ -9,8 +9,8 @@ from open_webui.models.knowledge import (
    KnowledgeResponse,
    KnowledgeUserResponse,
 )
-from open_webui.models.files import Files, FileModel
-from open_webui.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.models.files import Files, FileModel, FileMetadataResponse
+from open_webui.retrieval.vector.factory import VECTOR_DB_CLIENT
 from open_webui.routers.retrieval import (
    process_file,
    ProcessFileForm,
@@ -178,10 +178,26 @@ async def reindex_knowledge_files(request: Request, user=Depends(get_verified_us

    log.info(f"Starting reindexing for {len(knowledge_bases)} knowledge bases")

-    for knowledge_base in knowledge_bases:
-        try:
-            files = Files.get_files_by_ids(knowledge_base.data.get("file_ids", []))
+    deleted_knowledge_bases = []

+    for knowledge_base in knowledge_bases:
+        # -- Robust error handling for missing or invalid data
+        if not knowledge_base.data or not isinstance(knowledge_base.data, dict):
+            log.warning(
+                f"Knowledge base {knowledge_base.id} has no data or invalid data ({knowledge_base.data!r}). Deleting."
+            )
+            try:
+                Knowledges.delete_knowledge_by_id(id=knowledge_base.id)
+                deleted_knowledge_bases.append(knowledge_base.id)
+            except Exception as e:
+                log.error(
+                    f"Failed to delete invalid knowledge base {knowledge_base.id}: {e}"
+                )
+            continue
+
+        try:
+            file_ids = knowledge_base.data.get("file_ids", [])
+            files = Files.get_files_by_ids(file_ids)
            try:
                if VECTOR_DB_CLIENT.has_collection(collection_name=knowledge_base.id):
                    VECTOR_DB_CLIENT.delete_collection(
@@ -189,10 +205,7 @@ async def reindex_knowledge_files(request: Request, user=Depends(get_verified_us
                    )
            except Exception as e:
                log.error(f"Error deleting collection {knowledge_base.id}: {str(e)}")
-                raise HTTPException(
-                    status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-                    detail=f"Error deleting vector DB collection",
-                )
+                continue  # Skip, don't raise

            failed_files = []
            for file in files:
@@ -213,10 +226,8 @@ async def reindex_knowledge_files(request: Request, user=Depends(get_verified_us

        except Exception as e:
            log.error(f"Error processing knowledge base {knowledge_base.id}: {str(e)}")
-            raise HTTPException(
-                status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-                detail=f"Error processing knowledge base",
-            )
+            # Don't raise, just continue
+            continue

        if failed_files:
            log.warning(
@@ -225,7 +236,9 @@ async def reindex_knowledge_files(request: Request, user=Depends(get_verified_us
            for failed in failed_files:
                log.warning(f"File ID: {failed['file_id']}, Error: {failed['error']}")

-    log.info("Reindexing completed successfully")
+    log.info(
+        f"Reindexing completed. Deleted {len(deleted_knowledge_bases)} invalid knowledge bases: {deleted_knowledge_bases}"
+    )
    return True


@@ -235,7 +248,7 @@ async def reindex_knowledge_files(request: Request, user=Depends(get_verified_us


 class KnowledgeFilesResponse(KnowledgeResponse):
-    files: list[FileModel]
+    files: list[FileMetadataResponse]


@router.get("/{id}", response_model=Optional[KnowledgeFilesResponse])
@@ -251,7 +264,7 @@ async def get_knowledge_by_id(id: str, user=Depends(get_verified_user)):
        ):

            file_ids = knowledge.data.get("file_ids", []) if knowledge.data else []
-            files = Files.get_files_by_ids(file_ids)
+            files = Files.get_file_metadatas_by_ids(file_ids)

            return KnowledgeFilesResponse(
                **knowledge.model_dump(),
@@ -379,7 +392,7 @@ def add_file_to_knowledge_by_id(
            knowledge = Knowledges.update_knowledge_data_by_id(id=id, data=data)

            if knowledge:
-                files = Files.get_files_by_ids(file_ids)
+                files = Files.get_file_metadatas_by_ids(file_ids)

                return KnowledgeFilesResponse(
                    **knowledge.model_dump(),
@@ -456,7 +469,7 @@ def update_file_from_knowledge_by_id(
        data = knowledge.data or {}
        file_ids = data.get("file_ids", [])

-        files = Files.get_files_by_ids(file_ids)
+        files = Files.get_file_metadatas_by_ids(file_ids)

        return KnowledgeFilesResponse(
            **knowledge.model_dump(),
@@ -538,7 +551,7 @@ def remove_file_from_knowledge_by_id(
            knowledge = Knowledges.update_knowledge_data_by_id(id=id, data=data)

            if knowledge:
-                files = Files.get_files_by_ids(file_ids)
+                files = Files.get_file_metadatas_by_ids(file_ids)

                return KnowledgeFilesResponse(
                    **knowledge.model_dump(),
@@ -734,7 +747,7 @@ def add_files_to_knowledge_batch(
        error_details = [f"{err.file_id}: {err.error}" for err in result.errors]
        return KnowledgeFilesResponse(
            **knowledge.model_dump(),
-            files=Files.get_files_by_ids(existing_file_ids),
+            files=Files.get_file_metadatas_by_ids(existing_file_ids),
            warnings={
                "message": "Some files failed to process",
                "errors": error_details,
@@ -742,5 +755,6 @@ def add_files_to_knowledge_batch(
        )

    return KnowledgeFilesResponse(
-        **knowledge.model_dump(), files=Files.get_files_by_ids(existing_file_ids)
+        **knowledge.model_dump(),
+        files=Files.get_file_metadatas_by_ids(existing_file_ids),
    )
@@ -4,7 +4,7 @@ import logging
 from typing import Optional

 from open_webui.models.memories import Memories, MemoryModel
-from open_webui.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.retrieval.vector.factory import VECTOR_DB_CLIENT
 from open_webui.utils.auth import get_verified_user
 from open_webui.env import SRC_LOG_LEVELS

@@ -0,0 +1,218 @@
+import json
+import logging
+from typing import Optional
+
+
+from fastapi import APIRouter, Depends, HTTPException, Request, status, BackgroundTasks
+from pydantic import BaseModel
+
+from open_webui.models.users import Users, UserResponse
+from open_webui.models.notes import Notes, NoteModel, NoteForm, NoteUserResponse
+
+from open_webui.config import ENABLE_ADMIN_CHAT_ACCESS, ENABLE_ADMIN_EXPORT
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import SRC_LOG_LEVELS
+
+
+from open_webui.utils.auth import get_admin_user, get_verified_user
+from open_webui.utils.access_control import has_access, has_permission
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["MODELS"])
+
+router = APIRouter()
+
+############################
+# GetNotes
+############################
+
+
+@router.get("/", response_model=list[NoteUserResponse])
+async def get_notes(request: Request, user=Depends(get_verified_user)):
+
+    if user.role != "admin" and not has_permission(
+        user.id, "features.notes", request.app.state.config.USER_PERMISSIONS
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.UNAUTHORIZED,
+        )
+
+    notes = [
+        NoteUserResponse(
+            **{
+                **note.model_dump(),
+                "user": UserResponse(**Users.get_user_by_id(note.user_id).model_dump()),
+            }
+        )
+        for note in Notes.get_notes_by_user_id(user.id, "write")
+    ]
+
+    return notes
+
+
+@router.get("/list", response_model=list[NoteUserResponse])
+async def get_note_list(request: Request, user=Depends(get_verified_user)):
+
+    if user.role != "admin" and not has_permission(
+        user.id, "features.notes", request.app.state.config.USER_PERMISSIONS
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.UNAUTHORIZED,
+        )
+
+    notes = [
+        NoteUserResponse(
+            **{
+                **note.model_dump(),
+                "user": UserResponse(**Users.get_user_by_id(note.user_id).model_dump()),
+            }
+        )
+        for note in Notes.get_notes_by_user_id(user.id, "read")
+    ]
+
+    return notes
+
+
+############################
+# CreateNewNote
+############################
+
+
+@router.post("/create", response_model=Optional[NoteModel])
+async def create_new_note(
+    request: Request, form_data: NoteForm, user=Depends(get_verified_user)
+):
+
+    if user.role != "admin" and not has_permission(
+        user.id, "features.notes", request.app.state.config.USER_PERMISSIONS
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.UNAUTHORIZED,
+        )
+
+    try:
+        note = Notes.insert_new_note(form_data, user.id)
+        return note
+    except Exception as e:
+        log.exception(e)
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT()
+        )
+
+
+############################
+# GetNoteById
+############################
+
+
+@router.get("/{id}", response_model=Optional[NoteModel])
+async def get_note_by_id(request: Request, id: str, user=Depends(get_verified_user)):
+    if user.role != "admin" and not has_permission(
+        user.id, "features.notes", request.app.state.config.USER_PERMISSIONS
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.UNAUTHORIZED,
+        )
+
+    note = Notes.get_note_by_id(id)
+    if not note:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND, detail=ERROR_MESSAGES.NOT_FOUND
+        )
+
+    if (
+        user.role != "admin"
+        and user.id != note.user_id
+        and not has_access(user.id, type="read", access_control=note.access_control)
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT()
+        )
+
+    return note
+
+
+############################
+# UpdateNoteById
+############################
+
+
+@router.post("/{id}/update", response_model=Optional[NoteModel])
+async def update_note_by_id(
+    request: Request, id: str, form_data: NoteForm, user=Depends(get_verified_user)
+):
+    if user.role != "admin" and not has_permission(
+        user.id, "features.notes", request.app.state.config.USER_PERMISSIONS
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.UNAUTHORIZED,
+        )
+
+    note = Notes.get_note_by_id(id)
+    if not note:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND, detail=ERROR_MESSAGES.NOT_FOUND
+        )
+
+    if (
+        user.role != "admin"
+        and user.id != note.user_id
+        and not has_access(user.id, type="write", access_control=note.access_control)
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT()
+        )
+
+    try:
+        note = Notes.update_note_by_id(id, form_data)
+        return note
+    except Exception as e:
+        log.exception(e)
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT()
+        )
+
+
+############################
+# DeleteNoteById
+############################
+
+
+@router.delete("/{id}/delete", response_model=bool)
+async def delete_note_by_id(request: Request, id: str, user=Depends(get_verified_user)):
+    if user.role != "admin" and not has_permission(
+        user.id, "features.notes", request.app.state.config.USER_PERMISSIONS
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail=ERROR_MESSAGES.UNAUTHORIZED,
+        )
+
+    note = Notes.get_note_by_id(id)
+    if not note:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND, detail=ERROR_MESSAGES.NOT_FOUND
+        )
+
+    if (
+        user.role != "admin"
+        and user.id != note.user_id
+        and not has_access(user.id, type="write", access_control=note.access_control)
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT()
+        )
+
+    try:
+        note = Notes.delete_note_by_id(id)
+        return True
+    except Exception as e:
+        log.exception(e)
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT()
+        )
@@ -54,6 +54,7 @@ from open_webui.config import (
 from open_webui.env import (
    ENV,
    SRC_LOG_LEVELS,
+    AIOHTTP_CLIENT_SESSION_SSL,
    AIOHTTP_CLIENT_TIMEOUT,
    AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST,
    BYPASS_MODEL_ACCESS_CONTROL,
@@ -91,6 +92,7 @@ async def send_get_request(url, key=None, user: UserModel = None):
                        else {}
                    ),
                },
+                ssl=AIOHTTP_CLIENT_SESSION_SSL,
            ) as response:
                return await response.json()
    except Exception as e:
@@ -141,6 +143,7 @@ async def send_post_request(
                    else {}
                ),
            },
+            ssl=AIOHTTP_CLIENT_SESSION_SSL,
        )
        r.raise_for_status()

@@ -216,7 +219,8 @@ async def verify_connection(
    key = form_data.key

    async with aiohttp.ClientSession(
-        timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST)
+        trust_env=True,
+        timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST),
    ) as session:
        try:
            async with session.get(
@@ -234,6 +238,7 @@ async def verify_connection(
                        else {}
                    ),
                },
+                ssl=AIOHTTP_CLIENT_SESSION_SSL,
            ) as r:
                if r.status != 200:
                    detail = f"HTTP Error: {r.status}"
@@ -335,6 +340,8 @@ async def get_all_models(request: Request, user: UserModel = None):
                    ),  # Legacy support
                )

+                connection_type = api_config.get("connection_type", "local")
+
                prefix_id = api_config.get("prefix_id", None)
                tags = api_config.get("tags", [])
                model_ids = api_config.get("model_ids", [])
@@ -347,14 +354,16 @@ async def get_all_models(request: Request, user: UserModel = None):
                        )
                    )

-                if prefix_id:
-                    for model in response.get("models", []):
+                for model in response.get("models", []):
+                    if prefix_id:
                        model["model"] = f"{prefix_id}.{model['model']}"

-                if tags:
-                    for model in response.get("models", []):
+                    if tags:
                        model["tags"] = tags

+                    if connection_type:
+                        model["connection_type"] = connection_type
+
        def merge_models_lists(model_lists):
            merged_models = {}

@@ -878,8 +887,16 @@ async def embed(
            )

    url = request.app.state.config.OLLAMA_BASE_URLS[url_idx]
+    api_config = request.app.state.config.OLLAMA_API_CONFIGS.get(
+        str(url_idx),
+        request.app.state.config.OLLAMA_API_CONFIGS.get(url, {}),  # Legacy support
+    )
    key = get_api_key(url_idx, url, request.app.state.config.OLLAMA_API_CONFIGS)

+    prefix_id = api_config.get("prefix_id", None)
+    if prefix_id:
+        form_data.model = form_data.model.replace(f"{prefix_id}.", "")
+
    try:
        r = requests.request(
            method="POST",
@@ -957,8 +974,16 @@ async def embeddings(
            )

    url = request.app.state.config.OLLAMA_BASE_URLS[url_idx]
+    api_config = request.app.state.config.OLLAMA_API_CONFIGS.get(
+        str(url_idx),
+        request.app.state.config.OLLAMA_API_CONFIGS.get(url, {}),  # Legacy support
+    )
    key = get_api_key(url_idx, url, request.app.state.config.OLLAMA_API_CONFIGS)

+    prefix_id = api_config.get("prefix_id", None)
+    if prefix_id:
+        form_data.model = form_data.model.replace(f"{prefix_id}.", "")
+
    try:
        r = requests.request(
            method="POST",
@@ -1006,7 +1031,7 @@ class GenerateCompletionForm(BaseModel):
    prompt: str
    suffix: Optional[str] = None
    images: Optional[list[str]] = None
-    format: Optional[str] = None
+    format: Optional[Union[dict, str]] = None
    options: Optional[dict] = None
    system: Optional[str] = None
    template: Optional[str] = None
@@ -1482,7 +1507,9 @@ async def download_file_stream(
    timeout = aiohttp.ClientTimeout(total=600)  # Set the timeout

    async with aiohttp.ClientSession(timeout=timeout, trust_env=True) as session:
-        async with session.get(file_url, headers=headers) as response:
+        async with session.get(
+            file_url, headers=headers, ssl=AIOHTTP_CLIENT_SESSION_SSL
+        ) as response:
            total_size = int(response.headers.get("content-length", 0)) + current_size

            with open(file_path, "ab+") as file:
@@ -1497,7 +1524,8 @@ async def download_file_stream(

                if done:
                    file.seek(0)
-                    hashed = calculate_sha256(file)
+                    chunk_size = 1024 * 1024 * 2
+                    hashed = calculate_sha256(file, chunk_size)
                    file.seek(0)

                    url = f"{ollama_url}/api/blobs/sha256:{hashed}"
@@ -1561,7 +1589,9 @@ async def upload_model(
    if url_idx is None:
        url_idx = 0
    ollama_url = request.app.state.config.OLLAMA_BASE_URLS[url_idx]
-    file_path = os.path.join(UPLOAD_DIR, file.filename)
+
+    filename = os.path.basename(file.filename)
+    file_path = os.path.join(UPLOAD_DIR, filename)
    os.makedirs(UPLOAD_DIR, exist_ok=True)

    # --- P1: save file locally ---
@@ -1606,13 +1636,13 @@ async def upload_model(
                os.remove(file_path)

                # Create model in ollama
-                model_name, ext = os.path.splitext(file.filename)
+                model_name, ext = os.path.splitext(filename)
                log.info(f"Created Model: {model_name}")  # DEBUG

                create_payload = {
                    "model": model_name,
                    # Reference the file by its original name => the uploaded blob's digest
-                    "files": {file.filename: f"sha256:{file_hash}"},
+                    "files": {filename: f"sha256:{file_hash}"},
                }
                log.info(f"Model Payload: {create_payload}")  # DEBUG

@@ -1629,7 +1659,7 @@ async def upload_model(
                    done_msg = {
                        "done": True,
                        "blob": f"sha256:{file_hash}",
-                        "name": file.filename,
+                        "name": filename,
                        "model_created": model_name,
                    }
                    yield f"data: {json.dumps(done_msg)}\n\n"
@@ -21,6 +21,7 @@ from open_webui.config import (
    CACHE_DIR,
 )
 from open_webui.env import (
+    AIOHTTP_CLIENT_SESSION_SSL,
    AIOHTTP_CLIENT_TIMEOUT,
    AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST,
    ENABLE_FORWARD_USER_INFO_HEADERS,
@@ -74,6 +75,7 @@ async def send_get_request(url, key=None, user: UserModel = None):
                        else {}
                    ),
                },
+                ssl=AIOHTTP_CLIENT_SESSION_SSL,
            ) as response:
                return await response.json()
    except Exception as e:
@@ -92,20 +94,19 @@ async def cleanup_response(
        await session.close()


-def openai_o1_o3_handler(payload):
+def openai_o_series_handler(payload):
    """
-    Handle o1, o3 specific parameters
+    Handle "o" series specific parameters
    """
    if "max_tokens" in payload:
-        # Remove "max_tokens" from the payload
+        # Convert "max_tokens" to "max_completion_tokens" for all o-series models
        payload["max_completion_tokens"] = payload["max_tokens"]
        del payload["max_tokens"]

-    # Fix: o1 and o3 do not support the "system" role directly.
-    # For older models like "o1-mini" or "o1-preview", use role "user".
-    # For newer o1/o3 models, replace "system" with "developer".
+    # Handle system role conversion based on model type
    if payload["messages"][0]["role"] == "system":
        model_lower = payload["model"].lower()
+        # Legacy models use "user" role instead of "system"
        if model_lower.startswith("o1-mini") or model_lower.startswith("o1-preview"):
            payload["messages"][0]["role"] = "user"
        else:
@@ -352,21 +353,22 @@ async def get_all_models_responses(request: Request, user: UserModel) -> list:
                ),  # Legacy support
            )

+            connection_type = api_config.get("connection_type", "external")
            prefix_id = api_config.get("prefix_id", None)
            tags = api_config.get("tags", [])

-            if prefix_id:
-                for model in (
-                    response if isinstance(response, list) else response.get("data", [])
-                ):
+            for model in (
+                response if isinstance(response, list) else response.get("data", [])
+            ):
+                if prefix_id:
                    model["id"] = f"{prefix_id}.{model['id']}"

-            if tags:
-                for model in (
-                    response if isinstance(response, list) else response.get("data", [])
-                ):
+                if tags:
                    model["tags"] = tags

+                if connection_type:
+                    model["connection_type"] = connection_type
+
    log.debug(f"get_all_models:responses() {responses}")
    return responses

@@ -414,6 +416,7 @@ async def get_all_models(request: Request, user: UserModel) -> dict[str, list]:
                            "name": model.get("name", model["id"]),
                            "owned_by": "openai",
                            "openai": model,
+                            "connection_type": model.get("connection_type", "external"),
                            "urlIdx": idx,
                        }
                        for model in models
@@ -460,58 +463,74 @@ async def get_models(
        url = request.app.state.config.OPENAI_API_BASE_URLS[url_idx]
        key = request.app.state.config.OPENAI_API_KEYS[url_idx]

+        api_config = request.app.state.config.OPENAI_API_CONFIGS.get(
+            str(url_idx),
+            request.app.state.config.OPENAI_API_CONFIGS.get(url, {}),  # Legacy support
+        )
+
        r = None
        async with aiohttp.ClientSession(
-            timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST)
+            trust_env=True,
+            timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST),
        ) as session:
            try:
-                async with session.get(
-                    f"{url}/models",
-                    headers={
-                        "Authorization": f"Bearer {key}",
-                        "Content-Type": "application/json",
-                        **(
-                            {
-                                "X-OpenWebUI-User-Name": user.name,
-                                "X-OpenWebUI-User-Id": user.id,
-                                "X-OpenWebUI-User-Email": user.email,
-                                "X-OpenWebUI-User-Role": user.role,
-                            }
-                            if ENABLE_FORWARD_USER_INFO_HEADERS
-                            else {}
-                        ),
-                    },
-                ) as r:
-                    if r.status != 200:
-                        # Extract response error details if available
-                        error_detail = f"HTTP Error: {r.status}"
-                        res = await r.json()
-                        if "error" in res:
-                            error_detail = f"External Error: {res['error']}"
-                        raise Exception(error_detail)
+                headers = {
+                    "Content-Type": "application/json",
+                    **(
+                        {
+                            "X-OpenWebUI-User-Name": user.name,
+                            "X-OpenWebUI-User-Id": user.id,
+                            "X-OpenWebUI-User-Email": user.email,
+                            "X-OpenWebUI-User-Role": user.role,
+                        }
+                        if ENABLE_FORWARD_USER_INFO_HEADERS
+                        else {}
+                    ),
+                }

-                    response_data = await r.json()
+                if api_config.get("azure", False):
+                    models = {
+                        "data": api_config.get("model_ids", []) or [],
+                        "object": "list",
+                    }
+                else:
+                    headers["Authorization"] = f"Bearer {key}"

-                    # Check if we're calling OpenAI API based on the URL
-                    if "api.openai.com" in url:
-                        # Filter models according to the specified conditions
-                        response_data["data"] = [
-                            model
-                            for model in response_data.get("data", [])
-                            if not any(
-                                name in model["id"]
-                                for name in [
-                                    "babbage",
-                                    "dall-e",
-                                    "davinci",
-                                    "embedding",
-                                    "tts",
-                                    "whisper",
-                                ]
-                            )
-                        ]
+                    async with session.get(
+                        f"{url}/models",
+                        headers=headers,
+                        ssl=AIOHTTP_CLIENT_SESSION_SSL,
+                    ) as r:
+                        if r.status != 200:
+                            # Extract response error details if available
+                            error_detail = f"HTTP Error: {r.status}"
+                            res = await r.json()
+                            if "error" in res:
+                                error_detail = f"External Error: {res['error']}"
+                            raise Exception(error_detail)

-                    models = response_data
+                        response_data = await r.json()
+
+                        # Check if we're calling OpenAI API based on the URL
+                        if "api.openai.com" in url:
+                            # Filter models according to the specified conditions
+                            response_data["data"] = [
+                                model
+                                for model in response_data.get("data", [])
+                                if not any(
+                                    name in model["id"]
+                                    for name in [
+                                        "babbage",
+                                        "dall-e",
+                                        "davinci",
+                                        "embedding",
+                                        "tts",
+                                        "whisper",
+                                    ]
+                                )
+                            ]
+
+                        models = response_data
            except aiohttp.ClientError as e:
                # ClientError covers all aiohttp requests issues
                log.exception(f"Client error: {str(e)}")
@@ -533,6 +552,8 @@ class ConnectionVerificationForm(BaseModel):
    url: str
    key: str

+    config: Optional[dict] = None
+

@router.post("/verify")
 async def verify_connection(
@@ -541,37 +562,64 @@ async def verify_connection(
    url = form_data.url
    key = form_data.key

+    api_config = form_data.config or {}
+
    async with aiohttp.ClientSession(
-        timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST)
+        trust_env=True,
+        timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST),
    ) as session:
        try:
-            async with session.get(
-                f"{url}/models",
-                headers={
-                    "Authorization": f"Bearer {key}",
-                    "Content-Type": "application/json",
-                    **(
-                        {
-                            "X-OpenWebUI-User-Name": user.name,
-                            "X-OpenWebUI-User-Id": user.id,
-                            "X-OpenWebUI-User-Email": user.email,
-                            "X-OpenWebUI-User-Role": user.role,
-                        }
-                        if ENABLE_FORWARD_USER_INFO_HEADERS
-                        else {}
-                    ),
-                },
-            ) as r:
-                if r.status != 200:
-                    # Extract response error details if available
-                    error_detail = f"HTTP Error: {r.status}"
-                    res = await r.json()
-                    if "error" in res:
-                        error_detail = f"External Error: {res['error']}"
-                    raise Exception(error_detail)
+            headers = {
+                "Content-Type": "application/json",
+                **(
+                    {
+                        "X-OpenWebUI-User-Name": user.name,
+                        "X-OpenWebUI-User-Id": user.id,
+                        "X-OpenWebUI-User-Email": user.email,
+                        "X-OpenWebUI-User-Role": user.role,
+                    }
+                    if ENABLE_FORWARD_USER_INFO_HEADERS
+                    else {}
+                ),
+            }

-                response_data = await r.json()
-                return response_data
+            if api_config.get("azure", False):
+                headers["api-key"] = key
+                api_version = api_config.get("api_version", "") or "2023-03-15-preview"
+
+                async with session.get(
+                    url=f"{url}/openai/models?api-version={api_version}",
+                    headers=headers,
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
+                ) as r:
+                    if r.status != 200:
+                        # Extract response error details if available
+                        error_detail = f"HTTP Error: {r.status}"
+                        res = await r.json()
+                        if "error" in res:
+                            error_detail = f"External Error: {res['error']}"
+                        raise Exception(error_detail)
+
+                    response_data = await r.json()
+                    return response_data
+            else:
+                headers["Authorization"] = f"Bearer {key}"
+
+                async with session.get(
+                    f"{url}/models",
+                    headers=headers,
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
+                ) as r:
+                    if r.status != 200:
+                        # Extract response error details if available
+                        error_detail = f"HTTP Error: {r.status}"
+                        res = await r.json()
+                        if "error" in res:
+                            error_detail = f"External Error: {res['error']}"
+                        raise Exception(error_detail)
+
+                    response_data = await r.json()
+                    return response_data

        except aiohttp.ClientError as e:
            # ClientError covers all aiohttp requests issues
@@ -585,6 +633,63 @@ async def verify_connection(
            raise HTTPException(status_code=500, detail=error_detail)


+def convert_to_azure_payload(
+    url,
+    payload: dict,
+):
+    model = payload.get("model", "")
+
+    # Filter allowed parameters based on Azure OpenAI API
+    allowed_params = {
+        "messages",
+        "temperature",
+        "role",
+        "content",
+        "contentPart",
+        "contentPartImage",
+        "enhancements",
+        "dataSources",
+        "n",
+        "stream",
+        "stop",
+        "max_tokens",
+        "presence_penalty",
+        "frequency_penalty",
+        "logit_bias",
+        "user",
+        "function_call",
+        "functions",
+        "tools",
+        "tool_choice",
+        "top_p",
+        "log_probs",
+        "top_logprobs",
+        "response_format",
+        "seed",
+        "max_completion_tokens",
+    }
+
+    # Special handling for o-series models
+    if model.startswith("o") and model.endswith("-mini"):
+        # Convert max_tokens to max_completion_tokens for o-series models
+        if "max_tokens" in payload:
+            payload["max_completion_tokens"] = payload["max_tokens"]
+            del payload["max_tokens"]
+
+        # Remove temperature if not 1 for o-series models
+        if "temperature" in payload and payload["temperature"] != 1:
+            log.debug(
+                f"Removing temperature parameter for o-series model {model} as only default value (1) is supported"
+            )
+            del payload["temperature"]
+
+    # Filter out unsupported parameters
+    payload = {k: v for k, v in payload.items() if k in allowed_params}
+
+    url = f"{url}/openai/deployments/{model}"
+    return url, payload
+
+
@router.post("/chat/completions")
 async def generate_chat_completion(
    request: Request,
@@ -666,10 +771,10 @@ async def generate_chat_completion(
    url = request.app.state.config.OPENAI_API_BASE_URLS[idx]
    key = request.app.state.config.OPENAI_API_KEYS[idx]

-    # Fix: o1,o3 does not support the "max_tokens" parameter, Modify "max_tokens" to "max_completion_tokens"
-    is_o1_o3 = payload["model"].lower().startswith(("o1", "o3-"))
-    if is_o1_o3:
-        payload = openai_o1_o3_handler(payload)
+    # Check if model is from "o" series
+    is_o_series = payload["model"].lower().startswith(("o1", "o3", "o4"))
+    if is_o_series:
+        payload = openai_o_series_handler(payload)
    elif "api.openai.com" not in url:
        # Remove "max_completion_tokens" from the payload for backward compatibility
        if "max_completion_tokens" in payload:
@@ -685,6 +790,38 @@ async def generate_chat_completion(
            convert_logit_bias_input_to_json(payload["logit_bias"])
        )

+    headers = {
+        "Content-Type": "application/json",
+        **(
+            {
+                "HTTP-Referer": "https://openwebui.com/",
+                "X-Title": "Open WebUI",
+            }
+            if "openrouter.ai" in url
+            else {}
+        ),
+        **(
+            {
+                "X-OpenWebUI-User-Name": user.name,
+                "X-OpenWebUI-User-Id": user.id,
+                "X-OpenWebUI-User-Email": user.email,
+                "X-OpenWebUI-User-Role": user.role,
+            }
+            if ENABLE_FORWARD_USER_INFO_HEADERS
+            else {}
+        ),
+    }
+
+    if api_config.get("azure", False):
+        request_url, payload = convert_to_azure_payload(url, payload)
+        api_version = api_config.get("api_version", "") or "2023-03-15-preview"
+        headers["api-key"] = key
+        headers["api-version"] = api_version
+        request_url = f"{request_url}/chat/completions?api-version={api_version}"
+    else:
+        request_url = f"{url}/chat/completions"
+        headers["Authorization"] = f"Bearer {key}"
+
    payload = json.dumps(payload)

    r = None
@@ -699,30 +836,10 @@ async def generate_chat_completion(

        r = await session.request(
            method="POST",
-            url=f"{url}/chat/completions",
+            url=request_url,
            data=payload,
-            headers={
-                "Authorization": f"Bearer {key}",
-                "Content-Type": "application/json",
-                **(
-                    {
-                        "HTTP-Referer": "https://openwebui.com/",
-                        "X-Title": "Open WebUI",
-                    }
-                    if "openrouter.ai" in url
-                    else {}
-                ),
-                **(
-                    {
-                        "X-OpenWebUI-User-Name": user.name,
-                        "X-OpenWebUI-User-Id": user.id,
-                        "X-OpenWebUI-User-Email": user.email,
-                        "X-OpenWebUI-User-Role": user.role,
-                    }
-                    if ENABLE_FORWARD_USER_INFO_HEADERS
-                    else {}
-                ),
-            },
+            headers=headers,
+            ssl=AIOHTTP_CLIENT_SESSION_SSL,
        )

        # Check if response is SSE
@@ -777,31 +894,54 @@ async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
    idx = 0
    url = request.app.state.config.OPENAI_API_BASE_URLS[idx]
    key = request.app.state.config.OPENAI_API_KEYS[idx]
+    api_config = request.app.state.config.OPENAI_API_CONFIGS.get(
+        str(idx),
+        request.app.state.config.OPENAI_API_CONFIGS.get(
+            request.app.state.config.OPENAI_API_BASE_URLS[idx], {}
+        ),  # Legacy support
+    )

    r = None
    session = None
    streaming = False

    try:
+        headers = {
+            "Content-Type": "application/json",
+            **(
+                {
+                    "X-OpenWebUI-User-Name": user.name,
+                    "X-OpenWebUI-User-Id": user.id,
+                    "X-OpenWebUI-User-Email": user.email,
+                    "X-OpenWebUI-User-Role": user.role,
+                }
+                if ENABLE_FORWARD_USER_INFO_HEADERS
+                else {}
+            ),
+        }
+
+        if api_config.get("azure", False):
+            headers["api-key"] = key
+            headers["api-version"] = (
+                api_config.get("api_version", "") or "2023-03-15-preview"
+            )
+
+            payload = json.loads(body)
+            url, payload = convert_to_azure_payload(url, payload)
+            body = json.dumps(payload).encode()
+
+            request_url = f"{url}/{path}?api-version={api_config.get('api_version', '2023-03-15-preview')}"
+        else:
+            headers["Authorization"] = f"Bearer {key}"
+            request_url = f"{url}/{path}"
+
        session = aiohttp.ClientSession(trust_env=True)
        r = await session.request(
            method=request.method,
-            url=f"{url}/{path}",
+            url=request_url,
            data=body,
-            headers={
-                "Authorization": f"Bearer {key}",
-                "Content-Type": "application/json",
-                **(
-                    {
-                        "X-OpenWebUI-User-Name": user.name,
-                        "X-OpenWebUI-User-Id": user.id,
-                        "X-OpenWebUI-User-Email": user.email,
-                        "X-OpenWebUI-User-Role": user.role,
-                    }
-                    if ENABLE_FORWARD_USER_INFO_HEADERS
-                    else {}
-                ),
-            },
+            headers=headers,
+            ssl=AIOHTTP_CLIENT_SESSION_SSL,
        )
        r.raise_for_status()

@@ -18,7 +18,7 @@ from pydantic import BaseModel
 from starlette.responses import FileResponse
 from typing import Optional

-from open_webui.env import SRC_LOG_LEVELS
+from open_webui.env import SRC_LOG_LEVELS, AIOHTTP_CLIENT_SESSION_SSL
 from open_webui.config import CACHE_DIR
 from open_webui.constants import ERROR_MESSAGES

@@ -66,10 +66,13 @@ async def process_pipeline_inlet_filter(request, payload, user, models):
    if "pipeline" in model:
        sorted_filters.append(model)

-    async with aiohttp.ClientSession() as session:
+    async with aiohttp.ClientSession(trust_env=True) as session:
        for filter in sorted_filters:
            urlIdx = filter.get("urlIdx")
-            if urlIdx is None:
+
+            try:
+                urlIdx = int(urlIdx)
+            except:
                continue

            url = request.app.state.config.OPENAI_API_BASE_URLS[urlIdx]
@@ -89,6 +92,7 @@ async def process_pipeline_inlet_filter(request, payload, user, models):
                    f"{url}/{filter['id']}/filter/inlet",
                    headers=headers,
                    json=request_data,
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
                ) as response:
                    payload = await response.json()
                    response.raise_for_status()
@@ -115,10 +119,13 @@ async def process_pipeline_outlet_filter(request, payload, user, models):
    if "pipeline" in model:
        sorted_filters = [model] + sorted_filters

-    async with aiohttp.ClientSession() as session:
+    async with aiohttp.ClientSession(trust_env=True) as session:
        for filter in sorted_filters:
            urlIdx = filter.get("urlIdx")
-            if urlIdx is None:
+
+            try:
+                urlIdx = int(urlIdx)
+            except:
                continue

            url = request.app.state.config.OPENAI_API_BASE_URLS[urlIdx]
@@ -138,6 +145,7 @@ async def process_pipeline_outlet_filter(request, payload, user, models):
                    f"{url}/{filter['id']}/filter/outlet",
                    headers=headers,
                    json=request_data,
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
                ) as response:
                    payload = await response.json()
                    response.raise_for_status()
@@ -197,8 +205,10 @@ async def upload_pipeline(
    user=Depends(get_admin_user),
 ):
    log.info(f"upload_pipeline: urlIdx={urlIdx}, filename={file.filename}")
+    filename = os.path.basename(file.filename)
+
    # Check if the uploaded file is a python file
-    if not (file.filename and file.filename.endswith(".py")):
+    if not (filename and filename.endswith(".py")):
        raise HTTPException(
            status_code=status.HTTP_400_BAD_REQUEST,
            detail="Only Python (.py) files are allowed.",
@@ -206,7 +216,7 @@ async def upload_pipeline(

    upload_folder = f"{CACHE_DIR}/pipelines"
    os.makedirs(upload_folder, exist_ok=True)
-    file_path = os.path.join(upload_folder, file.filename)
+    file_path = os.path.join(upload_folder, filename)

    r = None
    try:
@@ -3,6 +3,8 @@ import logging
 import mimetypes
 import os
 import shutil
+import asyncio
+

 import uuid
 from datetime import datetime
@@ -34,7 +36,7 @@ from open_webui.models.knowledge import Knowledges
 from open_webui.storage.provider import Storage


-from open_webui.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.retrieval.vector.factory import VECTOR_DB_CLIENT

 # Document loaders
 from open_webui.retrieval.loaders.main import Loader
@@ -53,6 +55,7 @@ from open_webui.retrieval.web.jina_search import search_jina
 from open_webui.retrieval.web.searchapi import search_searchapi
 from open_webui.retrieval.web.serpapi import search_serpapi
 from open_webui.retrieval.web.searxng import search_searxng
+from open_webui.retrieval.web.yacy import search_yacy
 from open_webui.retrieval.web.serper import search_serper
 from open_webui.retrieval.web.serply import search_serply
 from open_webui.retrieval.web.serpstack import search_serpstack
@@ -61,6 +64,8 @@ from open_webui.retrieval.web.bing import search_bing
 from open_webui.retrieval.web.exa import search_exa
 from open_webui.retrieval.web.perplexity import search_perplexity
 from open_webui.retrieval.web.sougou import search_sougou
+from open_webui.retrieval.web.firecrawl import search_firecrawl
+from open_webui.retrieval.web.external import search_external

 from open_webui.retrieval.utils import (
    get_embedding_function,
@@ -90,7 +95,12 @@ from open_webui.env import (
    SRC_LOG_LEVELS,
    DEVICE_TYPE,
    DOCKER,
+    SENTENCE_TRANSFORMERS_BACKEND,
+    SENTENCE_TRANSFORMERS_MODEL_KWARGS,
+    SENTENCE_TRANSFORMERS_CROSS_ENCODER_BACKEND,
+    SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS,
 )
+
 from open_webui.constants import ERROR_MESSAGES

 log = logging.getLogger(__name__)
@@ -117,6 +127,8 @@ def get_ef(
                get_model_path(embedding_model, auto_update),
                device=DEVICE_TYPE,
                trust_remote_code=RAG_EMBEDDING_MODEL_TRUST_REMOTE_CODE,
+                backend=SENTENCE_TRANSFORMERS_BACKEND,
+                model_kwargs=SENTENCE_TRANSFORMERS_MODEL_KWARGS,
            )
        except Exception as e:
            log.debug(f"Error loading SentenceTransformer: {e}")
@@ -125,7 +137,10 @@ def get_ef(


 def get_rf(
+    engine: str = "",
    reranking_model: Optional[str] = None,
+    external_reranker_url: str = "",
+    external_reranker_api_key: str = "",
    auto_update: bool = False,
 ):
    rf = None
@@ -143,17 +158,33 @@ def get_rf(
                log.error(f"ColBERT: {e}")
                raise Exception(ERROR_MESSAGES.DEFAULT(e))
        else:
-            import sentence_transformers
+            if engine == "external":
+                try:
+                    from open_webui.retrieval.models.external import ExternalReranker
+
+                    rf = ExternalReranker(
+                        url=external_reranker_url,
+                        api_key=external_reranker_api_key,
+                        model=reranking_model,
+                    )
+                except Exception as e:
+                    log.error(f"ExternalReranking: {e}")
+                    raise Exception(ERROR_MESSAGES.DEFAULT(e))
+            else:
+                import sentence_transformers
+
+                try:
+                    rf = sentence_transformers.CrossEncoder(
+                        get_model_path(reranking_model, auto_update),
+                        device=DEVICE_TYPE,
+                        trust_remote_code=RAG_RERANKING_MODEL_TRUST_REMOTE_CODE,
+                        backend=SENTENCE_TRANSFORMERS_CROSS_ENCODER_BACKEND,
+                        model_kwargs=SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS,
+                    )
+                except Exception as e:
+                    log.error(f"CrossEncoder: {e}")
+                    raise Exception(ERROR_MESSAGES.DEFAULT("CrossEncoder error"))

-            try:
-                rf = sentence_transformers.CrossEncoder(
-                    get_model_path(reranking_model, auto_update),
-                    device=DEVICE_TYPE,
-                    trust_remote_code=RAG_RERANKING_MODEL_TRUST_REMOTE_CODE,
-                )
-            except Exception as e:
-                log.error(f"CrossEncoder: {e}")
-                raise Exception(ERROR_MESSAGES.DEFAULT("CrossEncoder error"))
    return rf


@@ -176,7 +207,7 @@ class ProcessUrlForm(CollectionNameForm):


 class SearchForm(BaseModel):
-    query: str
+    queries: List[str]


@router.get("/")
@@ -211,14 +242,6 @@ async def get_embedding_config(request: Request, user=Depends(get_admin_user)):
    }


-@router.get("/reranking")
-async def get_reraanking_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        "status": True,
-        "reranking_model": request.app.state.config.RAG_RERANKING_MODEL,
-    }
-
-
 class OpenAIConfigForm(BaseModel):
    url: str
    key: str
@@ -313,41 +336,6 @@ async def update_embedding_config(
        )


-class RerankingModelUpdateForm(BaseModel):
-    reranking_model: str
-
-
-@router.post("/reranking/update")
-async def update_reranking_config(
-    request: Request, form_data: RerankingModelUpdateForm, user=Depends(get_admin_user)
-):
-    log.info(
-        f"Updating reranking model: {request.app.state.config.RAG_RERANKING_MODEL} to {form_data.reranking_model}"
-    )
-    try:
-        request.app.state.config.RAG_RERANKING_MODEL = form_data.reranking_model
-
-        try:
-            request.app.state.rf = get_rf(
-                request.app.state.config.RAG_RERANKING_MODEL,
-                True,
-            )
-        except Exception as e:
-            log.error(f"Error loading reranking model: {e}")
-            request.app.state.config.ENABLE_RAG_HYBRID_SEARCH = False
-
-        return {
-            "status": True,
-            "reranking_model": request.app.state.config.RAG_RERANKING_MODEL,
-        }
-    except Exception as e:
-        log.exception(f"Problem updating reranking model: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail=ERROR_MESSAGES.DEFAULT(e),
-        )
-
-
@router.get("/config")
 async def get_rag_config(request: Request, user=Depends(get_admin_user)):
    return {
@@ -364,11 +352,21 @@ async def get_rag_config(request: Request, user=Depends(get_admin_user)):
        # Content extraction settings
        "CONTENT_EXTRACTION_ENGINE": request.app.state.config.CONTENT_EXTRACTION_ENGINE,
        "PDF_EXTRACT_IMAGES": request.app.state.config.PDF_EXTRACT_IMAGES,
+        "EXTERNAL_DOCUMENT_LOADER_URL": request.app.state.config.EXTERNAL_DOCUMENT_LOADER_URL,
+        "EXTERNAL_DOCUMENT_LOADER_API_KEY": request.app.state.config.EXTERNAL_DOCUMENT_LOADER_API_KEY,
        "TIKA_SERVER_URL": request.app.state.config.TIKA_SERVER_URL,
        "DOCLING_SERVER_URL": request.app.state.config.DOCLING_SERVER_URL,
+        "DOCLING_OCR_ENGINE": request.app.state.config.DOCLING_OCR_ENGINE,
+        "DOCLING_OCR_LANG": request.app.state.config.DOCLING_OCR_LANG,
+        "DOCLING_DO_PICTURE_DESCRIPTION": request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION,
        "DOCUMENT_INTELLIGENCE_ENDPOINT": request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT,
        "DOCUMENT_INTELLIGENCE_KEY": request.app.state.config.DOCUMENT_INTELLIGENCE_KEY,
        "MISTRAL_OCR_API_KEY": request.app.state.config.MISTRAL_OCR_API_KEY,
+        # Reranking settings
+        "RAG_RERANKING_MODEL": request.app.state.config.RAG_RERANKING_MODEL,
+        "RAG_RERANKING_ENGINE": request.app.state.config.RAG_RERANKING_ENGINE,
+        "RAG_EXTERNAL_RERANKER_URL": request.app.state.config.RAG_EXTERNAL_RERANKER_URL,
+        "RAG_EXTERNAL_RERANKER_API_KEY": request.app.state.config.RAG_EXTERNAL_RERANKER_API_KEY,
        # Chunking settings
        "TEXT_SPLITTER": request.app.state.config.TEXT_SPLITTER,
        "CHUNK_SIZE": request.app.state.config.CHUNK_SIZE,
@@ -376,6 +374,7 @@ async def get_rag_config(request: Request, user=Depends(get_admin_user)):
        # File upload settings
        "FILE_MAX_SIZE": request.app.state.config.FILE_MAX_SIZE,
        "FILE_MAX_COUNT": request.app.state.config.FILE_MAX_COUNT,
+        "ALLOWED_FILE_EXTENSIONS": request.app.state.config.ALLOWED_FILE_EXTENSIONS,
        # Integration settings
        "ENABLE_GOOGLE_DRIVE_INTEGRATION": request.app.state.config.ENABLE_GOOGLE_DRIVE_INTEGRATION,
        "ENABLE_ONEDRIVE_INTEGRATION": request.app.state.config.ENABLE_ONEDRIVE_INTEGRATION,
@@ -389,6 +388,9 @@ async def get_rag_config(request: Request, user=Depends(get_admin_user)):
            "WEB_SEARCH_DOMAIN_FILTER_LIST": request.app.state.config.WEB_SEARCH_DOMAIN_FILTER_LIST,
            "BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL": request.app.state.config.BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL,
            "SEARXNG_QUERY_URL": request.app.state.config.SEARXNG_QUERY_URL,
+            "YACY_QUERY_URL": request.app.state.config.YACY_QUERY_URL,
+            "YACY_USERNAME": request.app.state.config.YACY_USERNAME,
+            "YACY_PASSWORD": request.app.state.config.YACY_PASSWORD,
            "GOOGLE_PSE_API_KEY": request.app.state.config.GOOGLE_PSE_API_KEY,
            "GOOGLE_PSE_ENGINE_ID": request.app.state.config.GOOGLE_PSE_ENGINE_ID,
            "BRAVE_SEARCH_API_KEY": request.app.state.config.BRAVE_SEARCH_API_KEY,
@@ -418,6 +420,10 @@ async def get_rag_config(request: Request, user=Depends(get_admin_user)):
            "FIRECRAWL_API_KEY": request.app.state.config.FIRECRAWL_API_KEY,
            "FIRECRAWL_API_BASE_URL": request.app.state.config.FIRECRAWL_API_BASE_URL,
            "TAVILY_EXTRACT_DEPTH": request.app.state.config.TAVILY_EXTRACT_DEPTH,
+            "EXTERNAL_WEB_SEARCH_URL": request.app.state.config.EXTERNAL_WEB_SEARCH_URL,
+            "EXTERNAL_WEB_SEARCH_API_KEY": request.app.state.config.EXTERNAL_WEB_SEARCH_API_KEY,
+            "EXTERNAL_WEB_LOADER_URL": request.app.state.config.EXTERNAL_WEB_LOADER_URL,
+            "EXTERNAL_WEB_LOADER_API_KEY": request.app.state.config.EXTERNAL_WEB_LOADER_API_KEY,
            "YOUTUBE_LOADER_LANGUAGE": request.app.state.config.YOUTUBE_LOADER_LANGUAGE,
            "YOUTUBE_LOADER_PROXY_URL": request.app.state.config.YOUTUBE_LOADER_PROXY_URL,
            "YOUTUBE_LOADER_TRANSLATION": request.app.state.YOUTUBE_LOADER_TRANSLATION,
@@ -434,6 +440,9 @@ class WebConfig(BaseModel):
    WEB_SEARCH_DOMAIN_FILTER_LIST: Optional[List[str]] = []
    BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL: Optional[bool] = None
    SEARXNG_QUERY_URL: Optional[str] = None
+    YACY_QUERY_URL: Optional[str] = None
+    YACY_USERNAME: Optional[str] = None
+    YACY_PASSWORD: Optional[str] = None
    GOOGLE_PSE_API_KEY: Optional[str] = None
    GOOGLE_PSE_ENGINE_ID: Optional[str] = None
    BRAVE_SEARCH_API_KEY: Optional[str] = None
@@ -463,6 +472,10 @@ class WebConfig(BaseModel):
    FIRECRAWL_API_KEY: Optional[str] = None
    FIRECRAWL_API_BASE_URL: Optional[str] = None
    TAVILY_EXTRACT_DEPTH: Optional[str] = None
+    EXTERNAL_WEB_SEARCH_URL: Optional[str] = None
+    EXTERNAL_WEB_SEARCH_API_KEY: Optional[str] = None
+    EXTERNAL_WEB_LOADER_URL: Optional[str] = None
+    EXTERNAL_WEB_LOADER_API_KEY: Optional[str] = None
    YOUTUBE_LOADER_LANGUAGE: Optional[List[str]] = None
    YOUTUBE_LOADER_PROXY_URL: Optional[str] = None
    YOUTUBE_LOADER_TRANSLATION: Optional[str] = None
@@ -483,12 +496,24 @@ class ConfigForm(BaseModel):
    # Content extraction settings
    CONTENT_EXTRACTION_ENGINE: Optional[str] = None
    PDF_EXTRACT_IMAGES: Optional[bool] = None
+    EXTERNAL_DOCUMENT_LOADER_URL: Optional[str] = None
+    EXTERNAL_DOCUMENT_LOADER_API_KEY: Optional[str] = None
+
    TIKA_SERVER_URL: Optional[str] = None
    DOCLING_SERVER_URL: Optional[str] = None
+    DOCLING_OCR_ENGINE: Optional[str] = None
+    DOCLING_OCR_LANG: Optional[str] = None
+    DOCLING_DO_PICTURE_DESCRIPTION: Optional[bool] = None
    DOCUMENT_INTELLIGENCE_ENDPOINT: Optional[str] = None
    DOCUMENT_INTELLIGENCE_KEY: Optional[str] = None
    MISTRAL_OCR_API_KEY: Optional[str] = None

+    # Reranking settings
+    RAG_RERANKING_MODEL: Optional[str] = None
+    RAG_RERANKING_ENGINE: Optional[str] = None
+    RAG_EXTERNAL_RERANKER_URL: Optional[str] = None
+    RAG_EXTERNAL_RERANKER_API_KEY: Optional[str] = None
+
    # Chunking settings
    TEXT_SPLITTER: Optional[str] = None
    CHUNK_SIZE: Optional[int] = None
@@ -497,6 +522,7 @@ class ConfigForm(BaseModel):
    # File upload settings
    FILE_MAX_SIZE: Optional[int] = None
    FILE_MAX_COUNT: Optional[int] = None
+    ALLOWED_FILE_EXTENSIONS: Optional[List[str]] = None

    # Integration settings
    ENABLE_GOOGLE_DRIVE_INTEGRATION: Optional[bool] = None
@@ -564,6 +590,16 @@ async def update_rag_config(
        if form_data.PDF_EXTRACT_IMAGES is not None
        else request.app.state.config.PDF_EXTRACT_IMAGES
    )
+    request.app.state.config.EXTERNAL_DOCUMENT_LOADER_URL = (
+        form_data.EXTERNAL_DOCUMENT_LOADER_URL
+        if form_data.EXTERNAL_DOCUMENT_LOADER_URL is not None
+        else request.app.state.config.EXTERNAL_DOCUMENT_LOADER_URL
+    )
+    request.app.state.config.EXTERNAL_DOCUMENT_LOADER_API_KEY = (
+        form_data.EXTERNAL_DOCUMENT_LOADER_API_KEY
+        if form_data.EXTERNAL_DOCUMENT_LOADER_API_KEY is not None
+        else request.app.state.config.EXTERNAL_DOCUMENT_LOADER_API_KEY
+    )
    request.app.state.config.TIKA_SERVER_URL = (
        form_data.TIKA_SERVER_URL
        if form_data.TIKA_SERVER_URL is not None
@@ -574,6 +610,23 @@ async def update_rag_config(
        if form_data.DOCLING_SERVER_URL is not None
        else request.app.state.config.DOCLING_SERVER_URL
    )
+    request.app.state.config.DOCLING_OCR_ENGINE = (
+        form_data.DOCLING_OCR_ENGINE
+        if form_data.DOCLING_OCR_ENGINE is not None
+        else request.app.state.config.DOCLING_OCR_ENGINE
+    )
+    request.app.state.config.DOCLING_OCR_LANG = (
+        form_data.DOCLING_OCR_LANG
+        if form_data.DOCLING_OCR_LANG is not None
+        else request.app.state.config.DOCLING_OCR_LANG
+    )
+
+    request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION = (
+        form_data.DOCLING_DO_PICTURE_DESCRIPTION
+        if form_data.DOCLING_DO_PICTURE_DESCRIPTION is not None
+        else request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION
+    )
+
    request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT = (
        form_data.DOCUMENT_INTELLIGENCE_ENDPOINT
        if form_data.DOCUMENT_INTELLIGENCE_ENDPOINT is not None
@@ -590,6 +643,49 @@ async def update_rag_config(
        else request.app.state.config.MISTRAL_OCR_API_KEY
    )

+    # Reranking settings
+    request.app.state.config.RAG_RERANKING_ENGINE = (
+        form_data.RAG_RERANKING_ENGINE
+        if form_data.RAG_RERANKING_ENGINE is not None
+        else request.app.state.config.RAG_RERANKING_ENGINE
+    )
+
+    request.app.state.config.RAG_EXTERNAL_RERANKER_URL = (
+        form_data.RAG_EXTERNAL_RERANKER_URL
+        if form_data.RAG_EXTERNAL_RERANKER_URL is not None
+        else request.app.state.config.RAG_EXTERNAL_RERANKER_URL
+    )
+
+    request.app.state.config.RAG_EXTERNAL_RERANKER_API_KEY = (
+        form_data.RAG_EXTERNAL_RERANKER_API_KEY
+        if form_data.RAG_EXTERNAL_RERANKER_API_KEY is not None
+        else request.app.state.config.RAG_EXTERNAL_RERANKER_API_KEY
+    )
+
+    log.info(
+        f"Updating reranking model: {request.app.state.config.RAG_RERANKING_MODEL} to {form_data.RAG_RERANKING_MODEL}"
+    )
+    try:
+        request.app.state.config.RAG_RERANKING_MODEL = form_data.RAG_RERANKING_MODEL
+
+        try:
+            request.app.state.rf = get_rf(
+                request.app.state.config.RAG_RERANKING_ENGINE,
+                request.app.state.config.RAG_RERANKING_MODEL,
+                request.app.state.config.RAG_EXTERNAL_RERANKER_URL,
+                request.app.state.config.RAG_EXTERNAL_RERANKER_API_KEY,
+                True,
+            )
+        except Exception as e:
+            log.error(f"Error loading reranking model: {e}")
+            request.app.state.config.ENABLE_RAG_HYBRID_SEARCH = False
+    except Exception as e:
+        log.exception(f"Problem updating reranking model: {e}")
+        raise HTTPException(
+            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            detail=ERROR_MESSAGES.DEFAULT(e),
+        )
+
    # Chunking settings
    request.app.state.config.TEXT_SPLITTER = (
        form_data.TEXT_SPLITTER
@@ -618,6 +714,11 @@ async def update_rag_config(
        if form_data.FILE_MAX_COUNT is not None
        else request.app.state.config.FILE_MAX_COUNT
    )
+    request.app.state.config.ALLOWED_FILE_EXTENSIONS = (
+        form_data.ALLOWED_FILE_EXTENSIONS
+        if form_data.ALLOWED_FILE_EXTENSIONS is not None
+        else request.app.state.config.ALLOWED_FILE_EXTENSIONS
+    )

    # Integration settings
    request.app.state.config.ENABLE_GOOGLE_DRIVE_INTEGRATION = (
@@ -651,6 +752,9 @@ async def update_rag_config(
            form_data.web.BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL
        )
        request.app.state.config.SEARXNG_QUERY_URL = form_data.web.SEARXNG_QUERY_URL
+        request.app.state.config.YACY_QUERY_URL = form_data.web.YACY_QUERY_URL
+        request.app.state.config.YACY_USERNAME = form_data.web.YACY_USERNAME
+        request.app.state.config.YACY_PASSWORD = form_data.web.YACY_PASSWORD
        request.app.state.config.GOOGLE_PSE_API_KEY = form_data.web.GOOGLE_PSE_API_KEY
        request.app.state.config.GOOGLE_PSE_ENGINE_ID = (
            form_data.web.GOOGLE_PSE_ENGINE_ID
@@ -697,6 +801,18 @@ async def update_rag_config(
        request.app.state.config.FIRECRAWL_API_BASE_URL = (
            form_data.web.FIRECRAWL_API_BASE_URL
        )
+        request.app.state.config.EXTERNAL_WEB_SEARCH_URL = (
+            form_data.web.EXTERNAL_WEB_SEARCH_URL
+        )
+        request.app.state.config.EXTERNAL_WEB_SEARCH_API_KEY = (
+            form_data.web.EXTERNAL_WEB_SEARCH_API_KEY
+        )
+        request.app.state.config.EXTERNAL_WEB_LOADER_URL = (
+            form_data.web.EXTERNAL_WEB_LOADER_URL
+        )
+        request.app.state.config.EXTERNAL_WEB_LOADER_API_KEY = (
+            form_data.web.EXTERNAL_WEB_LOADER_API_KEY
+        )
        request.app.state.config.TAVILY_EXTRACT_DEPTH = (
            form_data.web.TAVILY_EXTRACT_DEPTH
        )
@@ -724,11 +840,21 @@ async def update_rag_config(
        # Content extraction settings
        "CONTENT_EXTRACTION_ENGINE": request.app.state.config.CONTENT_EXTRACTION_ENGINE,
        "PDF_EXTRACT_IMAGES": request.app.state.config.PDF_EXTRACT_IMAGES,
+        "EXTERNAL_DOCUMENT_LOADER_URL": request.app.state.config.EXTERNAL_DOCUMENT_LOADER_URL,
+        "EXTERNAL_DOCUMENT_LOADER_API_KEY": request.app.state.config.EXTERNAL_DOCUMENT_LOADER_API_KEY,
        "TIKA_SERVER_URL": request.app.state.config.TIKA_SERVER_URL,
        "DOCLING_SERVER_URL": request.app.state.config.DOCLING_SERVER_URL,
+        "DOCLING_OCR_ENGINE": request.app.state.config.DOCLING_OCR_ENGINE,
+        "DOCLING_OCR_LANG": request.app.state.config.DOCLING_OCR_LANG,
+        "DOCLING_DO_PICTURE_DESCRIPTION": request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION,
        "DOCUMENT_INTELLIGENCE_ENDPOINT": request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT,
        "DOCUMENT_INTELLIGENCE_KEY": request.app.state.config.DOCUMENT_INTELLIGENCE_KEY,
        "MISTRAL_OCR_API_KEY": request.app.state.config.MISTRAL_OCR_API_KEY,
+        # Reranking settings
+        "RAG_RERANKING_MODEL": request.app.state.config.RAG_RERANKING_MODEL,
+        "RAG_RERANKING_ENGINE": request.app.state.config.RAG_RERANKING_ENGINE,
+        "RAG_EXTERNAL_RERANKER_URL": request.app.state.config.RAG_EXTERNAL_RERANKER_URL,
+        "RAG_EXTERNAL_RERANKER_API_KEY": request.app.state.config.RAG_EXTERNAL_RERANKER_API_KEY,
        # Chunking settings
        "TEXT_SPLITTER": request.app.state.config.TEXT_SPLITTER,
        "CHUNK_SIZE": request.app.state.config.CHUNK_SIZE,
@@ -736,6 +862,7 @@ async def update_rag_config(
        # File upload settings
        "FILE_MAX_SIZE": request.app.state.config.FILE_MAX_SIZE,
        "FILE_MAX_COUNT": request.app.state.config.FILE_MAX_COUNT,
+        "ALLOWED_FILE_EXTENSIONS": request.app.state.config.ALLOWED_FILE_EXTENSIONS,
        # Integration settings
        "ENABLE_GOOGLE_DRIVE_INTEGRATION": request.app.state.config.ENABLE_GOOGLE_DRIVE_INTEGRATION,
        "ENABLE_ONEDRIVE_INTEGRATION": request.app.state.config.ENABLE_ONEDRIVE_INTEGRATION,
@@ -749,6 +876,9 @@ async def update_rag_config(
            "WEB_SEARCH_DOMAIN_FILTER_LIST": request.app.state.config.WEB_SEARCH_DOMAIN_FILTER_LIST,
            "BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL": request.app.state.config.BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL,
            "SEARXNG_QUERY_URL": request.app.state.config.SEARXNG_QUERY_URL,
+            "YACY_QUERY_URL": request.app.state.config.YACY_QUERY_URL,
+            "YACY_USERNAME": request.app.state.config.YACY_USERNAME,
+            "YACY_PASSWORD": request.app.state.config.YACY_PASSWORD,
            "GOOGLE_PSE_API_KEY": request.app.state.config.GOOGLE_PSE_API_KEY,
            "GOOGLE_PSE_ENGINE_ID": request.app.state.config.GOOGLE_PSE_ENGINE_ID,
            "BRAVE_SEARCH_API_KEY": request.app.state.config.BRAVE_SEARCH_API_KEY,
@@ -778,6 +908,10 @@ async def update_rag_config(
            "FIRECRAWL_API_KEY": request.app.state.config.FIRECRAWL_API_KEY,
            "FIRECRAWL_API_BASE_URL": request.app.state.config.FIRECRAWL_API_BASE_URL,
            "TAVILY_EXTRACT_DEPTH": request.app.state.config.TAVILY_EXTRACT_DEPTH,
+            "EXTERNAL_WEB_SEARCH_URL": request.app.state.config.EXTERNAL_WEB_SEARCH_URL,
+            "EXTERNAL_WEB_SEARCH_API_KEY": request.app.state.config.EXTERNAL_WEB_SEARCH_API_KEY,
+            "EXTERNAL_WEB_LOADER_URL": request.app.state.config.EXTERNAL_WEB_LOADER_URL,
+            "EXTERNAL_WEB_LOADER_API_KEY": request.app.state.config.EXTERNAL_WEB_LOADER_API_KEY,
            "YOUTUBE_LOADER_LANGUAGE": request.app.state.config.YOUTUBE_LOADER_LANGUAGE,
            "YOUTUBE_LOADER_PROXY_URL": request.app.state.config.YOUTUBE_LOADER_PROXY_URL,
            "YOUTUBE_LOADER_TRANSLATION": request.app.state.YOUTUBE_LOADER_TRANSLATION,
@@ -1030,8 +1164,13 @@ def process_file(
                file_path = Storage.get_file(file_path)
                loader = Loader(
                    engine=request.app.state.config.CONTENT_EXTRACTION_ENGINE,
+                    EXTERNAL_DOCUMENT_LOADER_URL=request.app.state.config.EXTERNAL_DOCUMENT_LOADER_URL,
+                    EXTERNAL_DOCUMENT_LOADER_API_KEY=request.app.state.config.EXTERNAL_DOCUMENT_LOADER_API_KEY,
                    TIKA_SERVER_URL=request.app.state.config.TIKA_SERVER_URL,
                    DOCLING_SERVER_URL=request.app.state.config.DOCLING_SERVER_URL,
+                    DOCLING_OCR_ENGINE=request.app.state.config.DOCLING_OCR_ENGINE,
+                    DOCLING_OCR_LANG=request.app.state.config.DOCLING_OCR_LANG,
+                    DOCLING_DO_PICTURE_DESCRIPTION=request.app.state.config.DOCLING_DO_PICTURE_DESCRIPTION,
                    PDF_EXTRACT_IMAGES=request.app.state.config.PDF_EXTRACT_IMAGES,
                    DOCUMENT_INTELLIGENCE_ENDPOINT=request.app.state.config.DOCUMENT_INTELLIGENCE_ENDPOINT,
                    DOCUMENT_INTELLIGENCE_KEY=request.app.state.config.DOCUMENT_INTELLIGENCE_KEY,
@@ -1266,6 +1405,7 @@ def search_web(request: Request, engine: str, query: str) -> list[SearchResult]:
    """Search the web using a search engine and return the results as a list of SearchResult objects.
    Will look for a search engine API key in environment variables in the following order:
    - SEARXNG_QUERY_URL
+    - YACY_QUERY_URL + YACY_USERNAME + YACY_PASSWORD
    - GOOGLE_PSE_API_KEY + GOOGLE_PSE_ENGINE_ID
    - BRAVE_SEARCH_API_KEY
    - KAGI_SEARCH_API_KEY
@@ -1295,6 +1435,18 @@ def search_web(request: Request, engine: str, query: str) -> list[SearchResult]:
            )
        else:
            raise Exception("No SEARXNG_QUERY_URL found in environment variables")
+    elif engine == "yacy":
+        if request.app.state.config.YACY_QUERY_URL:
+            return search_yacy(
+                request.app.state.config.YACY_QUERY_URL,
+                request.app.state.config.YACY_USERNAME,
+                request.app.state.config.YACY_PASSWORD,
+                query,
+                request.app.state.config.WEB_SEARCH_RESULT_COUNT,
+                request.app.state.config.WEB_SEARCH_DOMAIN_FILTER_LIST,
+            )
+        else:
+            raise Exception("No YACY_QUERY_URL found in environment variables")
    elif engine == "google_pse":
        if (
            request.app.state.config.GOOGLE_PSE_API_KEY
@@ -1465,6 +1617,22 @@ def search_web(request: Request, engine: str, query: str) -> list[SearchResult]:
            raise Exception(
                "No SOUGOU_API_SID or SOUGOU_API_SK found in environment variables"
            )
+    elif engine == "firecrawl":
+        return search_firecrawl(
+            request.app.state.config.FIRECRAWL_API_BASE_URL,
+            request.app.state.config.FIRECRAWL_API_KEY,
+            query,
+            request.app.state.config.WEB_SEARCH_RESULT_COUNT,
+            request.app.state.config.WEB_SEARCH_DOMAIN_FILTER_LIST,
+        )
+    elif engine == "external":
+        return search_external(
+            request.app.state.config.EXTERNAL_WEB_SEARCH_URL,
+            request.app.state.config.EXTERNAL_WEB_SEARCH_API_KEY,
+            query,
+            request.app.state.config.WEB_SEARCH_RESULT_COUNT,
+            request.app.state.config.WEB_SEARCH_DOMAIN_FILTER_LIST,
+        )
    else:
        raise Exception("No search engine API key found in environment variables")

@@ -1473,13 +1641,34 @@ def search_web(request: Request, engine: str, query: str) -> list[SearchResult]:
 async def process_web_search(
    request: Request, form_data: SearchForm, user=Depends(get_verified_user)
 ):
+
+    urls = []
    try:
        logging.info(
-            f"trying to web search with {request.app.state.config.WEB_SEARCH_ENGINE, form_data.query}"
-        )
-        web_results = search_web(
-            request, request.app.state.config.WEB_SEARCH_ENGINE, form_data.query
+            f"trying to web search with {request.app.state.config.WEB_SEARCH_ENGINE, form_data.queries}"
        )
+
+        search_tasks = [
+            run_in_threadpool(
+                search_web,
+                request,
+                request.app.state.config.WEB_SEARCH_ENGINE,
+                query,
+            )
+            for query in form_data.queries
+        ]
+
+        search_results = await asyncio.gather(*search_tasks)
+
+        for result in search_results:
+            if result:
+                for item in result:
+                    if item and item.link:
+                        urls.append(item.link)
+
+        urls = list(dict.fromkeys(urls))
+        log.debug(f"urls: {urls}")
+
    except Exception as e:
        log.exception(e)

@@ -1488,10 +1677,7 @@ async def process_web_search(
            detail=ERROR_MESSAGES.WEB_SEARCH_ERROR(e),
        )

-    log.debug(f"web_results: {web_results}")
-
    try:
-        urls = [result.link for result in web_results]
        loader = get_web_loader(
            urls,
            verify_ssl=request.app.state.config.ENABLE_WEB_LOADER_SSL_VERIFICATION,
@@ -1500,8 +1686,8 @@ async def process_web_search(
        )
        docs = await loader.aload()
        urls = [
-            doc.metadata["source"] for doc in docs
-        ]  # only keep URLs which could be retrieved
+            doc.metadata.get("source") for doc in docs if doc.metadata.get("source")
+        ]  # only keep the urls returned by the loader

        if request.app.state.config.BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL:
            return {
@@ -1518,26 +1704,28 @@ async def process_web_search(
                "loaded_count": len(docs),
            }
        else:
-            collection_names = []
-            for doc_idx, doc in enumerate(docs):
-                collection_name = f"web-search-{calculate_sha256_string(form_data.query + '-' + urls[doc_idx])}"[
+            # Create a single collection for all documents
+            collection_name = (
+                f"web-search-{calculate_sha256_string('-'.join(form_data.queries))}"[
                    :63
                ]
+            )

-                collection_names.append(collection_name)
-
+            try:
                await run_in_threadpool(
                    save_docs_to_vector_db,
                    request,
-                    [doc],
+                    docs,
                    collection_name,
                    overwrite=True,
                    user=user,
                )
+            except Exception as e:
+                log.debug(f"error saving docs: {e}")

            return {
                "status": True,
-                "collection_names": collection_names,
+                "collection_names": [collection_name],
                "filenames": urls,
                "loaded_count": len(docs),
            }
@@ -20,10 +20,7 @@ from open_webui.utils.auth import get_admin_user, get_verified_user
 from open_webui.constants import TASKS

 from open_webui.routers.pipelines import process_pipeline_inlet_filter
-from open_webui.utils.filter import (
-    get_sorted_filter_ids,
-    process_filter_functions,
-)
+
 from open_webui.utils.task import get_task_model_id

 from open_webui.config import (
@@ -186,35 +183,28 @@ async def generate_title(
    else:
        template = DEFAULT_TITLE_GENERATION_PROMPT_TEMPLATE

-    messages = form_data["messages"]
-
-    # Remove reasoning details from the messages
-    for message in messages:
-        message["content"] = re.sub(
-            r"<details\s+type=\"reasoning\"[^>]*>.*?<\/details>",
-            "",
-            message["content"],
-            flags=re.S,
-        ).strip()
-
    content = title_generation_template(
        template,
-        messages,
+        form_data["messages"],
        {
            "name": user.name,
            "location": user.info.get("location") if user.info else None,
        },
    )

+    max_tokens = (
+        models[task_model_id].get("info", {}).get("params", {}).get("max_tokens", 1000)
+    )
+
    payload = {
        "model": task_model_id,
        "messages": [{"role": "user", "content": content}],
        "stream": False,
        **(
-            {"max_tokens": 1000}
+            {"max_tokens": max_tokens}
            if models[task_model_id].get("owned_by") == "ollama"
            else {
-                "max_completion_tokens": 1000,
+                "max_completion_tokens": max_tokens,
            }
        ),
        "metadata": {
@@ -6,6 +6,7 @@ from open_webui.models.groups import Groups
 from open_webui.models.chats import Chats
 from open_webui.models.users import (
    UserModel,
+    UserListResponse,
    UserRoleUpdateForm,
    Users,
    UserSettings,
@@ -20,7 +21,7 @@ from fastapi import APIRouter, Depends, HTTPException, Request, status
 from pydantic import BaseModel

 from open_webui.utils.auth import get_admin_user, get_password_hash, get_verified_user
-from open_webui.utils.access_control import get_permissions
+from open_webui.utils.access_control import get_permissions, has_permission


 log = logging.getLogger(__name__)
@@ -33,13 +34,38 @@ router = APIRouter()
 ############################


-@router.get("/", response_model=list[UserModel])
+PAGE_ITEM_COUNT = 30
+
+
+@router.get("/", response_model=UserListResponse)
 async def get_users(
-    skip: Optional[int] = None,
-    limit: Optional[int] = None,
+    query: Optional[str] = None,
+    order_by: Optional[str] = None,
+    direction: Optional[str] = None,
+    page: Optional[int] = 1,
    user=Depends(get_admin_user),
 ):
-    return Users.get_users(skip, limit)
+    limit = PAGE_ITEM_COUNT
+
+    page = max(1, page)
+    skip = (page - 1) * limit
+
+    filter = {}
+    if query:
+        filter["query"] = query
+    if order_by:
+        filter["order_by"] = order_by
+    if direction:
+        filter["direction"] = direction
+
+    return Users.get_users(filter=filter, skip=skip, limit=limit)
+
+
+@router.get("/all", response_model=UserListResponse)
+async def get_all_users(
+    user=Depends(get_admin_user),
+):
+    return Users.get_users()


 ############################
@@ -88,6 +114,11 @@ class ChatPermissions(BaseModel):
    file_upload: bool = True
    delete: bool = True
    edit: bool = True
+    share: bool = True
+    export: bool = True
+    stt: bool = True
+    tts: bool = True
+    call: bool = True
    multiple_models: bool = True
    temporary: bool = True
    temporary_enforced: bool = False
@@ -98,6 +129,7 @@ class FeaturesPermissions(BaseModel):
    web_search: bool = True
    image_generation: bool = True
    code_interpreter: bool = True
+    notes: bool = True


 class UserPermissions(BaseModel):
@@ -173,9 +205,22 @@ async def get_user_settings_by_session_user(user=Depends(get_verified_user)):

@router.post("/user/settings/update", response_model=UserSettings)
 async def update_user_settings_by_session_user(
-    form_data: UserSettings, user=Depends(get_verified_user)
+    request: Request, form_data: UserSettings, user=Depends(get_verified_user)
 ):
-    user = Users.update_user_settings_by_id(user.id, form_data.model_dump())
+    updated_user_settings = form_data.model_dump()
+    if (
+        user.role != "admin"
+        and "toolServers" in updated_user_settings.get("ui").keys()
+        and not has_permission(
+            user.id,
+            "features.direct_tool_servers",
+            request.app.state.config.USER_PERMISSIONS,
+        )
+    ):
+        # If the user is not an admin and does not have permission to use tool servers, remove the key
+        updated_user_settings["ui"].pop("toolServers", None)
+
+    user = Users.update_user_settings_by_id(user.id, updated_user_settings)
    if user:
        return user.settings
    else:
@@ -285,6 +330,21 @@ async def update_user_by_id(
    form_data: UserUpdateForm,
    session_user=Depends(get_admin_user),
 ):
+    # Prevent modification of the primary admin user by other admins
+    try:
+        first_user = Users.get_first_user()
+        if first_user and user_id == first_user.id and session_user.id != user_id:
+            raise HTTPException(
+                status_code=status.HTTP_403_FORBIDDEN,
+                detail=ERROR_MESSAGES.ACTION_PROHIBITED,
+            )
+    except Exception as e:
+        log.error(f"Error checking primary admin status: {e}")
+        raise HTTPException(
+            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            detail="Could not verify primary admin status.",
+        )
+
    user = Users.get_user_by_id(user_id)

    if user:
@@ -332,6 +392,21 @@ async def update_user_by_id(

@router.delete("/{user_id}", response_model=bool)
 async def delete_user_by_id(user_id: str, user=Depends(get_admin_user)):
+    # Prevent deletion of the primary admin user
+    try:
+        first_user = Users.get_first_user()
+        if first_user and user_id == first_user.id:
+            raise HTTPException(
+                status_code=status.HTTP_403_FORBIDDEN,
+                detail=ERROR_MESSAGES.ACTION_PROHIBITED,
+            )
+    except Exception as e:
+        log.error(f"Error checking primary admin status: {e}")
+        raise HTTPException(
+            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            detail="Could not verify primary admin status.",
+        )
+
    if user.id != user_id:
        result = Auths.delete_auth_by_id(user_id)

@@ -343,6 +418,7 @@ async def delete_user_by_id(user_id: str, user=Depends(get_admin_user)):
            detail=ERROR_MESSAGES.DELETE_USER_ERROR,
        )

+    # Prevent self-deletion
    raise HTTPException(
        status_code=status.HTTP_403_FORBIDDEN,
        detail=ERROR_MESSAGES.ACTION_PROHIBITED,
@@ -159,18 +159,19 @@ def get_models_in_use():

@sio.on("usage")
 async def usage(sid, data):
-    model_id = data["model"]
-    # Record the timestamp for the last update
-    current_time = int(time.time())
+    if sid in SESSION_POOL:
+        model_id = data["model"]
+        # Record the timestamp for the last update
+        current_time = int(time.time())

-    # Store the new usage data and task
-    USAGE_POOL[model_id] = {
-        **(USAGE_POOL[model_id] if model_id in USAGE_POOL else {}),
-        sid: {"updated_at": current_time},
-    }
+        # Store the new usage data and task
+        USAGE_POOL[model_id] = {
+            **(USAGE_POOL[model_id] if model_id in USAGE_POOL else {}),
+            sid: {"updated_at": current_time},
+        }

-    # Broadcast the usage data to all clients
-    await sio.emit("usage", {"models": get_models_in_use()})
+        # Broadcast the usage data to all clients
+        await sio.emit("usage", {"models": get_models_in_use()})


@sio.event
@@ -278,7 +279,8 @@ async def channel_events(sid, data):

@sio.on("user-list")
 async def user_list(sid):
-    await sio.emit("user-list", {"user_ids": list(USER_POOL.keys())})
+    if sid in SESSION_POOL:
+        await sio.emit("user-list", {"user_ids": list(USER_POOL.keys())})


@sio.event
@@ -314,8 +316,8 @@ def get_event_emitter(request_info, update_db=True):
            )
        )

-        for session_id in session_ids:
-            await sio.emit(
+        emit_tasks = [
+            sio.emit(
                "chat-events",
                {
                    "chat_id": request_info.get("chat_id", None),
@@ -324,6 +326,10 @@ def get_event_emitter(request_info, update_db=True):
                },
                to=session_id,
            )
+            for session_id in session_ids
+        ]
+
+        await asyncio.gather(*emit_tasks)

        if update_db:
            if "type" in event_data and event_data["type"] == "status":
@@ -339,16 +345,17 @@ def get_event_emitter(request_info, update_db=True):
                    request_info["message_id"],
                )

-                content = message.get("content", "")
-                content += event_data.get("data", {}).get("content", "")
+                if message:
+                    content = message.get("content", "")
+                    content += event_data.get("data", {}).get("content", "")

-                Chats.upsert_message_to_chat_by_id_and_message_id(
-                    request_info["chat_id"],
-                    request_info["message_id"],
-                    {
-                        "content": content,
-                    },
-                )
+                    Chats.upsert_message_to_chat_by_id_and_message_id(
+                        request_info["chat_id"],
+                        request_info["message_id"],
+                        {
+                            "content": content,
+                        },
+                    )

            if "type" in event_data and event_data["type"] == "replace":
                content = event_data.get("data", {}).get("content", "")
@@ -3,7 +3,7 @@ import shutil
 import json
 import logging
 from abc import ABC, abstractmethod
-from typing import BinaryIO, Tuple
+from typing import BinaryIO, Tuple, Dict

 import boto3
 from botocore.config import Config
@@ -17,6 +17,7 @@ from open_webui.config import (
    S3_SECRET_ACCESS_KEY,
    S3_USE_ACCELERATE_ENDPOINT,
    S3_ADDRESSING_STYLE,
+    S3_ENABLE_TAGGING,
    GCS_BUCKET_NAME,
    GOOGLE_APPLICATION_CREDENTIALS_JSON,
    AZURE_STORAGE_ENDPOINT,
@@ -44,7 +45,9 @@ class StorageProvider(ABC):
        pass

    @abstractmethod
-    def upload_file(self, file: BinaryIO, filename: str) -> Tuple[bytes, str]:
+    def upload_file(
+        self, file: BinaryIO, filename: str, tags: Dict[str, str]
+    ) -> Tuple[bytes, str]:
        pass

    @abstractmethod
@@ -58,7 +61,9 @@ class StorageProvider(ABC):

 class LocalStorageProvider(StorageProvider):
    @staticmethod
-    def upload_file(file: BinaryIO, filename: str) -> Tuple[bytes, str]:
+    def upload_file(
+        file: BinaryIO, filename: str, tags: Dict[str, str]
+    ) -> Tuple[bytes, str]:
        contents = file.read()
        if not contents:
            raise ValueError(ERROR_MESSAGES.EMPTY_CONTENT)
@@ -131,15 +136,24 @@ class S3StorageProvider(StorageProvider):
        self.bucket_name = S3_BUCKET_NAME
        self.key_prefix = S3_KEY_PREFIX if S3_KEY_PREFIX else ""

-    def upload_file(self, file: BinaryIO, filename: str) -> Tuple[bytes, str]:
+    def upload_file(
+        self, file: BinaryIO, filename: str, tags: Dict[str, str]
+    ) -> Tuple[bytes, str]:
        """Handles uploading of the file to S3 storage."""
-        _, file_path = LocalStorageProvider.upload_file(file, filename)
+        _, file_path = LocalStorageProvider.upload_file(file, filename, tags)
+        s3_key = os.path.join(self.key_prefix, filename)
        try:
-            s3_key = os.path.join(self.key_prefix, filename)
            self.s3_client.upload_file(file_path, self.bucket_name, s3_key)
+            if S3_ENABLE_TAGGING and tags:
+                tagging = {"TagSet": [{"Key": k, "Value": v} for k, v in tags.items()]}
+                self.s3_client.put_object_tagging(
+                    Bucket=self.bucket_name,
+                    Key=s3_key,
+                    Tagging=tagging,
+                )
            return (
                open(file_path, "rb").read(),
-                "s3://" + self.bucket_name + "/" + s3_key,
+                f"s3://{self.bucket_name}/{s3_key}",
            )
        except ClientError as e:
            raise RuntimeError(f"Error uploading file to S3: {e}")
@@ -207,9 +221,11 @@ class GCSStorageProvider(StorageProvider):
            self.gcs_client = storage.Client()
        self.bucket = self.gcs_client.bucket(GCS_BUCKET_NAME)

-    def upload_file(self, file: BinaryIO, filename: str) -> Tuple[bytes, str]:
+    def upload_file(
+        self, file: BinaryIO, filename: str, tags: Dict[str, str]
+    ) -> Tuple[bytes, str]:
        """Handles uploading of the file to GCS storage."""
-        contents, file_path = LocalStorageProvider.upload_file(file, filename)
+        contents, file_path = LocalStorageProvider.upload_file(file, filename, tags)
        try:
            blob = self.bucket.blob(filename)
            blob.upload_from_filename(file_path)
@@ -277,9 +293,11 @@ class AzureStorageProvider(StorageProvider):
            self.container_name
        )

-    def upload_file(self, file: BinaryIO, filename: str) -> Tuple[bytes, str]:
+    def upload_file(
+        self, file: BinaryIO, filename: str, tags: Dict[str, str]
+    ) -> Tuple[bytes, str]:
        """Handles uploading of the file to Azure Blob Storage."""
-        contents, file_path = LocalStorageProvider.upload_file(file, filename)
+        contents, file_path = LocalStorageProvider.upload_file(file, filename, tags)
        try:
            blob_client = self.container_client.get_blob_client(filename)
            blob_client.upload_blob(contents, overwrite=True)
@@ -37,7 +37,7 @@ if TYPE_CHECKING:
 class AuditLogEntry:
    # `Metadata` audit level properties
    id: str
-    user: dict[str, Any]
+    user: Optional[dict[str, Any]]
    audit_level: str
    verb: str
    request_uri: str
@@ -190,21 +190,40 @@ class AuditLoggingMiddleware:
        finally:
            await self._log_audit_entry(request, context)

-    async def _get_authenticated_user(self, request: Request) -> UserModel:
-
+    async def _get_authenticated_user(self, request: Request) -> Optional[UserModel]:
        auth_header = request.headers.get("Authorization")
-        assert auth_header
-        user = get_current_user(request, None, get_http_authorization_cred(auth_header))

-        return user
+        try:
+            user = get_current_user(
+                request, None, get_http_authorization_cred(auth_header)
+            )
+            return user
+        except Exception as e:
+            logger.debug(f"Failed to get authenticated user: {str(e)}")
+
+        return None

    def _should_skip_auditing(self, request: Request) -> bool:
        if (
            request.method not in {"POST", "PUT", "PATCH", "DELETE"}
            or AUDIT_LOG_LEVEL == "NONE"
-            or not request.headers.get("authorization")
        ):
            return True
+
+        ALWAYS_LOG_ENDPOINTS = {
+            "/api/v1/auths/signin",
+            "/api/v1/auths/signout",
+            "/api/v1/auths/signup",
+        }
+        path = request.url.path.lower()
+        for endpoint in ALWAYS_LOG_ENDPOINTS:
+            if path.startswith(endpoint):
+                return False  # Do NOT skip logging for auth endpoints
+
+        # Skip logging if the request is not authenticated
+        if not request.headers.get("authorization"):
+            return True
+
        # match either /api/<resource>/...(for the endpoint /api/chat case) or /api/v1/<resource>/...
        pattern = re.compile(
            r"^/api(?:/v1)?/(" + "|".join(self.excluded_paths) + r")\b"
@@ -231,17 +250,32 @@ class AuditLoggingMiddleware:
        try:
            user = await self._get_authenticated_user(request)

+            user = (
+                user.model_dump(include={"id", "name", "email", "role"}) if user else {}
+            )
+
+            request_body = context.request_body.decode("utf-8", errors="replace")
+            response_body = context.response_body.decode("utf-8", errors="replace")
+
+            # Redact sensitive information
+            if "password" in request_body:
+                request_body = re.sub(
+                    r'"password":\s*"(.*?)"',
+                    '"password": "********"',
+                    request_body,
+                )
+
            entry = AuditLogEntry(
                id=str(uuid.uuid4()),
-                user=user.model_dump(include={"id", "name", "email", "role"}),
+                user=user,
                audit_level=self.audit_level.value,
                verb=request.method,
                request_uri=str(request.url),
                response_status_code=context.metadata.get("response_status_code", None),
                source_ip=request.client.host if request.client else None,
                user_agent=request.headers.get("user-agent"),
-                request_object=context.request_body.decode("utf-8", errors="replace"),
-                response_object=context.response_body.decode("utf-8", errors="replace"),
+                request_object=request_body,
+                response_object=response_body,
            )

            self.audit_logger.write(entry)
@@ -13,6 +13,8 @@ import pytz
 from pytz import UTC
 from typing import Optional, Union, List, Dict

+from opentelemetry import trace
+
 from open_webui.models.users import Users

 from open_webui.constants import ERROR_MESSAGES
@@ -194,7 +196,17 @@ def get_current_user(
                    status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.API_KEY_NOT_ALLOWED
                )

-        return get_current_user_by_api_key(token)
+        user = get_current_user_by_api_key(token)
+
+        # Add user info to current span
+        current_span = trace.get_current_span()
+        if current_span:
+            current_span.set_attribute("client.user.id", user.id)
+            current_span.set_attribute("client.user.email", user.email)
+            current_span.set_attribute("client.user.role", user.role)
+            current_span.set_attribute("client.auth.type", "api_key")
+
+        return user

    # auth by jwt token
    try:
@@ -213,6 +225,14 @@ def get_current_user(
                detail=ERROR_MESSAGES.INVALID_TOKEN,
            )
        else:
+            # Add user info to current span
+            current_span = trace.get_current_span()
+            if current_span:
+                current_span.set_attribute("client.user.id", user.id)
+                current_span.set_attribute("client.user.email", user.email)
+                current_span.set_attribute("client.user.role", user.role)
+                current_span.set_attribute("client.auth.type", "jwt")
+
            # Refresh the user's last active timestamp asynchronously
            # to prevent blocking the request
            if background_tasks:
@@ -234,6 +254,14 @@ def get_current_user_by_api_key(api_key: str):
            detail=ERROR_MESSAGES.INVALID_TOKEN,
        )
    else:
+        # Add user info to current span
+        current_span = trace.get_current_span()
+        if current_span:
+            current_span.set_attribute("client.user.id", user.id)
+            current_span.set_attribute("client.user.email", user.email)
+            current_span.set_attribute("client.user.role", user.role)
+            current_span.set_attribute("client.auth.type", "api_key")
+
        Users.update_user_last_active_by_id(user.id)

    return user
@@ -309,6 +309,7 @@ async def chat_completed(request: Request, form_data: dict, user: Any):
    metadata = {
        "chat_id": data["chat_id"],
        "message_id": data["id"],
+        "filter_ids": data.get("filter_ids", []),
        "session_id": data["session_id"],
        "user_id": user.id,
    }
@@ -330,7 +331,9 @@ async def chat_completed(request: Request, form_data: dict, user: Any):
    try:
        filter_functions = [
            Functions.get_function_by_id(filter_id)
-            for filter_id in get_sorted_filter_ids(model)
+            for filter_id in get_sorted_filter_ids(
+                request, model, metadata.get("filter_ids", [])
+            )
        ]

        result, _ = await process_filter_functions(
@@ -44,13 +44,15 @@ class JupyterCodeExecuter:
        :param password: Jupyter password (optional)
        :param timeout: WebSocket timeout in seconds (default: 60s)
        """
-        self.base_url = base_url.rstrip("/")
+        self.base_url = base_url
        self.code = code
        self.token = token
        self.password = password
        self.timeout = timeout
        self.kernel_id = ""
-        self.session = aiohttp.ClientSession(base_url=self.base_url)
+        if self.base_url[-1] != "/":
+            self.base_url += "/"
+        self.session = aiohttp.ClientSession(trust_env=True, base_url=self.base_url)
        self.params = {}
        self.result = ResultModel()

@@ -61,7 +63,7 @@ class JupyterCodeExecuter:
        if self.kernel_id:
            try:
                async with self.session.delete(
-                    f"/api/kernels/{self.kernel_id}", params=self.params
+                    f"api/kernels/{self.kernel_id}", params=self.params
                ) as response:
                    response.raise_for_status()
            except Exception as err:
@@ -81,7 +83,7 @@ class JupyterCodeExecuter:
    async def sign_in(self) -> None:
        # password authentication
        if self.password and not self.token:
-            async with self.session.get("/login") as response:
+            async with self.session.get("login") as response:
                response.raise_for_status()
                xsrf_token = response.cookies["_xsrf"].value
                if not xsrf_token:
@@ -89,7 +91,7 @@ class JupyterCodeExecuter:
                self.session.cookie_jar.update_cookies(response.cookies)
                self.session.headers.update({"X-XSRFToken": xsrf_token})
            async with self.session.post(
-                "/login",
+                "login",
                data={"_xsrf": xsrf_token, "password": self.password},
                allow_redirects=False,
            ) as response:
@@ -101,17 +103,15 @@ class JupyterCodeExecuter:
            self.params.update({"token": self.token})

    async def init_kernel(self) -> None:
-        async with self.session.post(
-            url="/api/kernels", params=self.params
-        ) as response:
+        async with self.session.post(url="api/kernels", params=self.params) as response:
            response.raise_for_status()
            kernel_data = await response.json()
            self.kernel_id = kernel_data["id"]

    def init_ws(self) -> (str, dict):
-        ws_base = self.base_url.replace("http", "ws")
+        ws_base = self.base_url.replace("http", "ws", 1)
        ws_params = "?" + "&".join([f"{key}={val}" for key, val in self.params.items()])
-        websocket_url = f"{ws_base}/api/kernels/{self.kernel_id}/channels{ws_params if len(ws_params) > 1 else ''}"
+        websocket_url = f"{ws_base}api/kernels/{self.kernel_id}/channels{ws_params if len(ws_params) > 1 else ''}"
        ws_headers = {}
        if self.password and not self.token:
            ws_headers = {
@@ -9,7 +9,20 @@ log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["MAIN"])


-def get_sorted_filter_ids(model: dict):
+def get_function_module(request, function_id):
+    """
+    Get the function module by its ID.
+    """
+    if function_id in request.app.state.FUNCTIONS:
+        function_module = request.app.state.FUNCTIONS[function_id]
+    else:
+        function_module, _, _ = load_function_module_by_id(function_id)
+        request.app.state.FUNCTIONS[function_id] = function_module
+
+    return function_module
+
+
+def get_sorted_filter_ids(request, model: dict, enabled_filter_ids: list = None):
    def get_priority(function_id):
        function = Functions.get_function_by_id(function_id)
        if function is not None:
@@ -21,14 +34,23 @@ def get_sorted_filter_ids(model: dict):
    if "info" in model and "meta" in model["info"]:
        filter_ids.extend(model["info"]["meta"].get("filterIds", []))
        filter_ids = list(set(filter_ids))
-
-    enabled_filter_ids = [
+    active_filter_ids = [
        function.id
        for function in Functions.get_functions_by_type("filter", active_only=True)
    ]

-    filter_ids = [fid for fid in filter_ids if fid in enabled_filter_ids]
+    for filter_id in active_filter_ids:
+        function_module = get_function_module(request, filter_id)
+
+        if getattr(function_module, "toggle", None) and (
+            filter_id not in enabled_filter_ids
+        ):
+            active_filter_ids.remove(filter_id)
+            continue
+
+    filter_ids = [fid for fid in filter_ids if fid in active_filter_ids]
    filter_ids.sort(key=get_priority)
+
    return filter_ids


@@ -43,12 +65,7 @@ async def process_filter_functions(
        if not filter:
            continue

-        if filter_id in request.app.state.FUNCTIONS:
-            function_module = request.app.state.FUNCTIONS[filter_id]
-        else:
-            function_module, _, _ = load_function_module_by_id(filter_id)
-            request.app.state.FUNCTIONS[filter_id] = function_module
-
+        function_module = get_function_module(request, filter_id)
        # Prepare handler function
        handler = getattr(function_module, filter_type, None)
        if not handler:
@@ -340,6 +340,11 @@ async def chat_web_search_handler(
        log.exception(e)
        queries = [user_message]

+    # Check if generated queries are empty
+    if len(queries) == 1 and queries[0].strip() == "":
+        queries = [user_message]
+
+    # Check if queries are not found
    if len(queries) == 0:
        await event_emitter(
            {
@@ -353,115 +358,86 @@ async def chat_web_search_handler(
        )
        return form_data

-    all_results = []
+    await event_emitter(
+        {
+            "type": "status",
+            "data": {
+                "action": "web_search",
+                "description": "Searching the web",
+                "done": False,
+            },
+        }
+    )

-    for searchQuery in queries:
-        await event_emitter(
-            {
-                "type": "status",
-                "data": {
-                    "action": "web_search",
-                    "description": 'Searching "{{searchQuery}}"',
-                    "query": searchQuery,
-                    "done": False,
-                },
-            }
+    try:
+        results = await process_web_search(
+            request,
+            SearchForm(queries=queries),
+            user=user,
        )

-        try:
-            results = await process_web_search(
-                request,
-                SearchForm(
-                    **{
-                        "query": searchQuery,
+        if results:
+            files = form_data.get("files", [])
+
+            if results.get("collection_names"):
+                for col_idx, collection_name in enumerate(
+                    results.get("collection_names")
+                ):
+                    files.append(
+                        {
+                            "collection_name": collection_name,
+                            "name": ", ".join(queries),
+                            "type": "web_search",
+                            "urls": results["filenames"],
+                        }
+                    )
+            elif results.get("docs"):
+                # Invoked when bypass embedding and retrieval is set to True
+                docs = results["docs"]
+                files.append(
+                    {
+                        "docs": docs,
+                        "name": ", ".join(queries),
+                        "type": "web_search",
+                        "urls": results["filenames"],
                    }
-                ),
-                user=user,
-            )
+                )

-            if results:
-                all_results.append(results)
-                files = form_data.get("files", [])
+            form_data["files"] = files

-                if results.get("collection_names"):
-                    for col_idx, collection_name in enumerate(
-                        results.get("collection_names")
-                    ):
-                        files.append(
-                            {
-                                "collection_name": collection_name,
-                                "name": searchQuery,
-                                "type": "web_search",
-                                "urls": [results["filenames"][col_idx]],
-                            }
-                        )
-                elif results.get("docs"):
-                    # Invoked when bypass embedding and retrieval is set to True
-                    docs = results["docs"]
-
-                    if len(docs) == len(results["filenames"]):
-                        # the number of docs and filenames (urls) should be the same
-                        for doc_idx, doc in enumerate(docs):
-                            files.append(
-                                {
-                                    "docs": [doc],
-                                    "name": searchQuery,
-                                    "type": "web_search",
-                                    "urls": [results["filenames"][doc_idx]],
-                                }
-                            )
-                    else:
-                        # edge case when the number of docs and filenames (urls) are not the same
-                        # this should not happen, but if it does, we will just append the docs
-                        files.append(
-                            {
-                                "docs": results.get("docs", []),
-                                "name": searchQuery,
-                                "type": "web_search",
-                                "urls": results["filenames"],
-                            }
-                        )
-
-                form_data["files"] = files
-        except Exception as e:
-            log.exception(e)
            await event_emitter(
                {
                    "type": "status",
                    "data": {
                        "action": "web_search",
-                        "description": 'Error searching "{{searchQuery}}"',
-                        "query": searchQuery,
+                        "description": "Searched {{count}} sites",
+                        "urls": results["filenames"],
+                        "done": True,
+                    },
+                }
+            )
+        else:
+            await event_emitter(
+                {
+                    "type": "status",
+                    "data": {
+                        "action": "web_search",
+                        "description": "No search results found",
                        "done": True,
                        "error": True,
                    },
                }
            )

-    if all_results:
-        urls = []
-        for results in all_results:
-            if "filenames" in results:
-                urls.extend(results["filenames"])
-
+    except Exception as e:
+        log.exception(e)
        await event_emitter(
            {
                "type": "status",
                "data": {
                    "action": "web_search",
-                    "description": "Searched {{count}} sites",
-                    "urls": urls,
-                    "done": True,
-                },
-            }
-        )
-    else:
-        await event_emitter(
-            {
-                "type": "status",
-                "data": {
-                    "action": "web_search",
-                    "description": "No search results found",
+                    "description": "An error occurred while searching the web",
+                    "queries": queries,
                    "done": True,
                    "error": True,
                },
@@ -534,13 +510,20 @@ async def chat_image_generation_handler(
            }
        )

-        for image in images:
-            await __event_emitter__(
-                {
-                    "type": "message",
-                    "data": {"content": f"![Generated Image]({image['url']})\n"},
-                }
-            )
+        await __event_emitter__(
+            {
+                "type": "files",
+                "data": {
+                    "files": [
+                        {
+                            "type": "image",
+                            "url": image["url"],
+                        }
+                        for image in images
+                    ]
+                },
+            }
+        )

        system_message_content = "<context>User is shown the generated image, tell the user that the image has been generated</context>"
    except Exception as e:
@@ -661,6 +644,9 @@ def apply_params_to_form_data(form_data, model):
        if "frequency_penalty" in params and params["frequency_penalty"] is not None:
            form_data["frequency_penalty"] = params["frequency_penalty"]

+        if "presence_penalty" in params and params["presence_penalty"] is not None:
+            form_data["presence_penalty"] = params["presence_penalty"]
+
        if "reasoning_effort" in params and params["reasoning_effort"] is not None:
            form_data["reasoning_effort"] = params["reasoning_effort"]

@@ -670,7 +656,7 @@ def apply_params_to_form_data(form_data, model):
                    convert_logit_bias_input_to_json(params["logit_bias"])
                )
            except Exception as e:
-                print(f"Error parsing logit_bias: {e}")
+                log.exception(f"Error parsing logit_bias: {e}")

    return form_data

@@ -768,9 +754,12 @@ async def process_chat_payload(request, form_data, user, metadata, model):
        raise e

    try:
+
        filter_functions = [
            Functions.get_function_by_id(filter_id)
-            for filter_id in get_sorted_filter_ids(model)
+            for filter_id in get_sorted_filter_ids(
+                request, model, metadata.get("filter_ids", [])
+            )
        ]

        form_data, flags = await process_filter_functions(
@@ -881,16 +870,20 @@ async def process_chat_payload(request, form_data, user, metadata, model):
    # If context is not empty, insert it into the messages
    if len(sources) > 0:
        context_string = ""
-        citated_file_idx = {}
-        for _, source in enumerate(sources, 1):
+        citation_idx = {}
+        for source in sources:
            if "document" in source:
                for doc_context, doc_meta in zip(
                    source["document"], source["metadata"]
                ):
-                    file_id = doc_meta.get("file_id")
-                    if file_id not in citated_file_idx:
-                        citated_file_idx[file_id] = len(citated_file_idx) + 1
-                    context_string += f'<source id="{citated_file_idx[file_id]}">{doc_context}</source>\n'
+                    citation_id = (
+                        doc_meta.get("source", None)
+                        or source.get("source", {}).get("id", None)
+                        or "N/A"
+                    )
+                    if citation_id not in citation_idx:
+                        citation_idx[citation_id] = len(citation_idx) + 1
+                    context_string += f'<source id="{citation_idx[citation_id]}">{doc_context}</source>\n'

        context_string = context_string.strip()
        prompt = get_last_user_message(form_data["messages"])
@@ -923,7 +916,12 @@ async def process_chat_payload(request, form_data, user, metadata, model):
            )

    # If there are citations, add them to the data_items
-    sources = [source for source in sources if source.get("source", {}).get("name", "")]
+    sources = [
+        source
+        for source in sources
+        if source.get("source", {}).get("name", "")
+        or source.get("source", {}).get("id", "")
+    ]

    if len(sources) > 0:
        events.append({"sources": sources})
@@ -952,7 +950,35 @@ async def process_chat_response(
        message = message_map.get(metadata["message_id"]) if message_map else None

        if message:
-            messages = get_message_list(message_map, message.get("id"))
+            message_list = get_message_list(message_map, message.get("id"))
+
+            # Remove details tags and files from the messages.
+            # as get_message_list creates a new list, it does not affect
+            # the original messages outside of this handler
+
+            messages = []
+            for message in message_list:
+                content = message.get("content", "")
+                if isinstance(content, list):
+                    for item in content:
+                        if item.get("type") == "text":
+                            content = item["text"]
+                            break
+
+                if isinstance(content, str):
+                    content = re.sub(
+                        r"<details\b[^>]*>.*?<\/details>",
+                        "",
+                        content,
+                        flags=re.S | re.I,
+                    ).strip()
+
+                messages.append(
+                    {
+                        "role": message["role"],
+                        "content": content,
+                    }
+                )

            if tasks and messages:
                if TASKS.TITLE_GENERATION in tasks:
@@ -1122,7 +1148,7 @@ async def process_chat_response(
                    )

                    # Send a webhook notification if the user is not active
-                    if get_active_status_by_user_id(user.id) is None:
+                    if not get_active_status_by_user_id(user.id):
                        webhook_url = Users.get_user_webhook_url_by_id(user.id)
                        if webhook_url:
                            post_webhook(
@@ -1165,7 +1191,9 @@ async def process_chat_response(
    }
    filter_functions = [
        Functions.get_function_by_id(filter_id)
-        for filter_id in get_sorted_filter_ids(model)
+        for filter_id in get_sorted_filter_ids(
+            request, model, metadata.get("filter_ids", [])
+        )
    ]

    # Streaming response
@@ -1410,6 +1438,9 @@ async def process_chat_response(

                            if after_tag:
                                content_blocks[-1]["content"] = after_tag
+                                tag_content_handler(
+                                    content_type, tags, after_tag, content_blocks
+                                )

                            break
                elif content_blocks[-1]["type"] == content_type:
@@ -1660,6 +1691,15 @@ async def process_chat_response(

                                                if current_response_tool_call is None:
                                                    # Add the new tool call
+                                                    delta_tool_call.setdefault(
+                                                        "function", {}
+                                                    )
+                                                    delta_tool_call[
+                                                        "function"
+                                                    ].setdefault("name", "")
+                                                    delta_tool_call[
+                                                        "function"
+                                                    ].setdefault("arguments", "")
                                                    response_tool_calls.append(
                                                        delta_tool_call
                                                    )
@@ -2204,7 +2244,7 @@ async def process_chat_response(
                    )

                # Send a webhook notification if the user is not active
-                if get_active_status_by_user_id(user.id) is None:
+                if not get_active_status_by_user_id(user.id):
                    webhook_url = Users.get_user_webhook_url_by_id(user.id)
                    if webhook_url:
                        post_webhook(
@@ -49,6 +49,7 @@ async def get_all_base_models(request: Request, user: UserModel = None):
                "created": int(time.time()),
                "owned_by": "ollama",
                "ollama": model,
+                "connection_type": model.get("connection_type", "local"),
                "tags": model.get("tags", []),
            }
            for model in ollama_models["models"]
@@ -110,6 +111,14 @@ async def get_all_models(request, user: UserModel = None):
        for function in Functions.get_functions_by_type("action", active_only=True)
    ]

+    global_filter_ids = [
+        function.id for function in Functions.get_global_filter_functions()
+    ]
+    enabled_filter_ids = [
+        function.id
+        for function in Functions.get_functions_by_type("filter", active_only=True)
+    ]
+
    custom_models = Models.get_all_models()
    for custom_model in custom_models:
        if custom_model.base_model_id is None:
@@ -125,13 +134,20 @@ async def get_all_models(request, user: UserModel = None):
                        model["name"] = custom_model.name
                        model["info"] = custom_model.model_dump()

+                        # Set action_ids and filter_ids
                        action_ids = []
+                        filter_ids = []
+
                        if "info" in model and "meta" in model["info"]:
                            action_ids.extend(
                                model["info"]["meta"].get("actionIds", [])
                            )
+                            filter_ids.extend(
+                                model["info"]["meta"].get("filterIds", [])
+                            )

                        model["action_ids"] = action_ids
+                        model["filter_ids"] = filter_ids
                    else:
                        models.remove(model)

@@ -140,7 +156,9 @@ async def get_all_models(request, user: UserModel = None):
        ):
            owned_by = "openai"
            pipe = None
+
            action_ids = []
+            filter_ids = []

            for model in models:
                if (
@@ -154,9 +172,13 @@ async def get_all_models(request, user: UserModel = None):

            if custom_model.meta:
                meta = custom_model.meta.model_dump()
+
                if "actionIds" in meta:
                    action_ids.extend(meta["actionIds"])

+                if "filterIds" in meta:
+                    filter_ids.extend(meta["filterIds"])
+
            models.append(
                {
                    "id": f"{custom_model.id}",
@@ -168,6 +190,7 @@ async def get_all_models(request, user: UserModel = None):
                    "preset": True,
                    **({"pipe": pipe} if pipe is not None else {}),
                    "action_ids": action_ids,
+                    "filter_ids": filter_ids,
                }
            )

@@ -181,8 +204,11 @@ async def get_all_models(request, user: UserModel = None):
                    "id": f"{function.id}.{action['id']}",
                    "name": action.get("name", f"{function.name} ({action['id']})"),
                    "description": function.meta.description,
-                    "icon_url": action.get(
-                        "icon_url", function.meta.manifest.get("icon_url", None)
+                    "icon": action.get(
+                        "icon_url",
+                        function.meta.manifest.get("icon_url", None)
+                        or getattr(module, "icon_url", None)
+                        or getattr(module, "icon", None),
                    ),
                }
                for action in actions
@@ -193,16 +219,32 @@ async def get_all_models(request, user: UserModel = None):
                    "id": function.id,
                    "name": function.name,
                    "description": function.meta.description,
-                    "icon_url": function.meta.manifest.get("icon_url", None),
+                    "icon": function.meta.manifest.get("icon_url", None)
+                    or getattr(module, "icon_url", None)
+                    or getattr(module, "icon", None),
                }
            ]

+    # Process filter_ids to get the filters
+    def get_filter_items_from_module(function, module):
+        return [
+            {
+                "id": function.id,
+                "name": function.name,
+                "description": function.meta.description,
+                "icon": function.meta.manifest.get("icon_url", None)
+                or getattr(module, "icon_url", None)
+                or getattr(module, "icon", None),
+            }
+        ]
+
    def get_function_module_by_id(function_id):
        if function_id in request.app.state.FUNCTIONS:
            function_module = request.app.state.FUNCTIONS[function_id]
        else:
            function_module, _, _ = load_function_module_by_id(function_id)
            request.app.state.FUNCTIONS[function_id] = function_module
+        return function_module

    for model in models:
        action_ids = [
@@ -210,6 +252,11 @@ async def get_all_models(request, user: UserModel = None):
            for action_id in list(set(model.pop("action_ids", []) + global_action_ids))
            if action_id in enabled_action_ids
        ]
+        filter_ids = [
+            filter_id
+            for filter_id in list(set(model.pop("filter_ids", []) + global_filter_ids))
+            if filter_id in enabled_filter_ids
+        ]

        model["actions"] = []
        for action_id in action_ids:
@@ -221,6 +268,20 @@ async def get_all_models(request, user: UserModel = None):
            model["actions"].extend(
                get_action_items_from_module(action_function, function_module)
            )
+
+        model["filters"] = []
+        for filter_id in filter_ids:
+            filter_function = Functions.get_function_by_id(filter_id)
+            if filter_function is None:
+                raise Exception(f"Filter not found: {filter_id}")
+
+            function_module = get_function_module_by_id(filter_id)
+
+            if getattr(function_module, "toggle", None):
+                model["filters"].extend(
+                    get_filter_items_from_module(filter_function, function_module)
+                )
+
    log.debug(f"get_all_models() returned {len(models)} models")

    request.app.state.MODELS = {model["id"]: model for model in models}
@@ -3,6 +3,7 @@ import logging
 import mimetypes
 import sys
 import uuid
+import json

 import aiohttp
 from authlib.integrations.starlette_client import OAuth
@@ -15,7 +16,7 @@ from starlette.responses import RedirectResponse

 from open_webui.models.auths import Auths
 from open_webui.models.users import Users
-from open_webui.models.groups import Groups, GroupModel, GroupUpdateForm
+from open_webui.models.groups import Groups, GroupModel, GroupUpdateForm, GroupForm
 from open_webui.config import (
    DEFAULT_USER_ROLE,
    ENABLE_OAUTH_SIGNUP,
@@ -23,6 +24,8 @@ from open_webui.config import (
    OAUTH_PROVIDERS,
    ENABLE_OAUTH_ROLE_MANAGEMENT,
    ENABLE_OAUTH_GROUP_MANAGEMENT,
+    ENABLE_OAUTH_GROUP_CREATION,
+    OAUTH_BLOCKED_GROUPS,
    OAUTH_ROLES_CLAIM,
    OAUTH_GROUPS_CLAIM,
    OAUTH_EMAIL_CLAIM,
@@ -31,12 +34,14 @@ from open_webui.config import (
    OAUTH_ALLOWED_ROLES,
    OAUTH_ADMIN_ROLES,
    OAUTH_ALLOWED_DOMAINS,
+    OAUTH_UPDATE_PICTURE_ON_LOGIN,
    WEBHOOK_URL,
    JWT_EXPIRES_IN,
    AppConfig,
 )
 from open_webui.constants import ERROR_MESSAGES, WEBHOOK_MESSAGES
 from open_webui.env import (
+    AIOHTTP_CLIENT_SESSION_SSL,
    WEBUI_NAME,
    WEBUI_AUTH_COOKIE_SAME_SITE,
    WEBUI_AUTH_COOKIE_SECURE,
@@ -57,6 +62,8 @@ auth_manager_config.ENABLE_OAUTH_SIGNUP = ENABLE_OAUTH_SIGNUP
 auth_manager_config.OAUTH_MERGE_ACCOUNTS_BY_EMAIL = OAUTH_MERGE_ACCOUNTS_BY_EMAIL
 auth_manager_config.ENABLE_OAUTH_ROLE_MANAGEMENT = ENABLE_OAUTH_ROLE_MANAGEMENT
 auth_manager_config.ENABLE_OAUTH_GROUP_MANAGEMENT = ENABLE_OAUTH_GROUP_MANAGEMENT
+auth_manager_config.ENABLE_OAUTH_GROUP_CREATION = ENABLE_OAUTH_GROUP_CREATION
+auth_manager_config.OAUTH_BLOCKED_GROUPS = OAUTH_BLOCKED_GROUPS
 auth_manager_config.OAUTH_ROLES_CLAIM = OAUTH_ROLES_CLAIM
 auth_manager_config.OAUTH_GROUPS_CLAIM = OAUTH_GROUPS_CLAIM
 auth_manager_config.OAUTH_EMAIL_CLAIM = OAUTH_EMAIL_CLAIM
@@ -67,6 +74,7 @@ auth_manager_config.OAUTH_ADMIN_ROLES = OAUTH_ADMIN_ROLES
 auth_manager_config.OAUTH_ALLOWED_DOMAINS = OAUTH_ALLOWED_DOMAINS
 auth_manager_config.WEBHOOK_URL = WEBHOOK_URL
 auth_manager_config.JWT_EXPIRES_IN = JWT_EXPIRES_IN
+auth_manager_config.OAUTH_UPDATE_PICTURE_ON_LOGIN = OAUTH_UPDATE_PICTURE_ON_LOGIN


 class OAuthManager:
@@ -140,6 +148,12 @@ class OAuthManager:
        log.debug("Running OAUTH Group management")
        oauth_claim = auth_manager_config.OAUTH_GROUPS_CLAIM

+        try:
+            blocked_groups = json.loads(auth_manager_config.OAUTH_BLOCKED_GROUPS)
+        except Exception as e:
+            log.exception(f"Error loading OAUTH_BLOCKED_GROUPS: {e}")
+            blocked_groups = []
+
        user_oauth_groups = []
        # Nested claim search for groups claim
        if oauth_claim:
@@ -147,11 +161,62 @@ class OAuthManager:
            nested_claims = oauth_claim.split(".")
            for nested_claim in nested_claims:
                claim_data = claim_data.get(nested_claim, {})
-            user_oauth_groups = claim_data if isinstance(claim_data, list) else []
+
+            if isinstance(claim_data, list):
+                user_oauth_groups = claim_data
+            elif isinstance(claim_data, str):
+                user_oauth_groups = [claim_data]
+            else:
+                user_oauth_groups = []

        user_current_groups: list[GroupModel] = Groups.get_groups_by_member_id(user.id)
        all_available_groups: list[GroupModel] = Groups.get_groups()

+        # Create groups if they don't exist and creation is enabled
+        if auth_manager_config.ENABLE_OAUTH_GROUP_CREATION:
+            log.debug("Checking for missing groups to create...")
+            all_group_names = {g.name for g in all_available_groups}
+            groups_created = False
+            # Determine creator ID: Prefer admin, fallback to current user if no admin exists
+            admin_user = Users.get_super_admin_user()
+            creator_id = admin_user.id if admin_user else user.id
+            log.debug(f"Using creator ID {creator_id} for potential group creation.")
+
+            for group_name in user_oauth_groups:
+                if group_name not in all_group_names:
+                    log.info(
+                        f"Group '{group_name}' not found via OAuth claim. Creating group..."
+                    )
+                    try:
+                        new_group_form = GroupForm(
+                            name=group_name,
+                            description=f"Group '{group_name}' created automatically via OAuth.",
+                            permissions=default_permissions,  # Use default permissions from function args
+                            user_ids=[],  # Start with no users, user will be added later by subsequent logic
+                        )
+                        # Use determined creator ID (admin or fallback to current user)
+                        created_group = Groups.insert_new_group(
+                            creator_id, new_group_form
+                        )
+                        if created_group:
+                            log.info(
+                                f"Successfully created group '{group_name}' with ID {created_group.id} using creator ID {creator_id}"
+                            )
+                            groups_created = True
+                            # Add to local set to prevent duplicate creation attempts in this run
+                            all_group_names.add(group_name)
+                        else:
+                            log.error(
+                                f"Failed to create group '{group_name}' via OAuth."
+                            )
+                    except Exception as e:
+                        log.error(f"Error creating group '{group_name}' via OAuth: {e}")
+
+            # Refresh the list of all available groups if any were created
+            if groups_created:
+                all_available_groups = Groups.get_groups()
+                log.debug("Refreshed list of all available groups after creation.")
+
        log.debug(f"Oauth Groups claim: {oauth_claim}")
        log.debug(f"User oauth groups: {user_oauth_groups}")
        log.debug(f"User's current groups: {[g.name for g in user_current_groups]}")
@@ -161,7 +226,11 @@ class OAuthManager:

        # Remove groups that user is no longer a part of
        for group_model in user_current_groups:
-            if user_oauth_groups and group_model.name not in user_oauth_groups:
+            if (
+                user_oauth_groups
+                and group_model.name not in user_oauth_groups
+                and group_model.name not in blocked_groups
+            ):
                # Remove group from user
                log.debug(
                    f"Removing user from group {group_model.name} as it is no longer in their oauth groups"
@@ -191,6 +260,7 @@ class OAuthManager:
                user_oauth_groups
                and group_model.name in user_oauth_groups
                and not any(gm.name == group_model.name for gm in user_current_groups)
+                and group_model.name not in blocked_groups
            ):
                # Add user to group
                log.debug(
@@ -215,6 +285,51 @@ class OAuthManager:
                    id=group_model.id, form_data=update_form, overwrite=False
                )

+    async def _process_picture_url(
+        self, picture_url: str, access_token: str = None
+    ) -> str:
+        """Process a picture URL and return a base64 encoded data URL.
+
+        Args:
+            picture_url: The URL of the picture to process
+            access_token: Optional OAuth access token for authenticated requests
+
+        Returns:
+            A data URL containing the base64 encoded picture, or "/user.png" if processing fails
+        """
+        if not picture_url:
+            return "/user.png"
+
+        try:
+            get_kwargs = {}
+            if access_token:
+                get_kwargs["headers"] = {
+                    "Authorization": f"Bearer {access_token}",
+                }
+            async with aiohttp.ClientSession(trust_env=True) as session:
+                async with session.get(
+                    picture_url, **get_kwargs, ssl=AIOHTTP_CLIENT_SESSION_SSL
+                ) as resp:
+                    if resp.ok:
+                        picture = await resp.read()
+                        base64_encoded_picture = base64.b64encode(picture).decode(
+                            "utf-8"
+                        )
+                        guessed_mime_type = mimetypes.guess_type(picture_url)[0]
+                        if guessed_mime_type is None:
+                            guessed_mime_type = "image/jpeg"
+                        return (
+                            f"data:{guessed_mime_type};base64,{base64_encoded_picture}"
+                        )
+                    else:
+                        log.warning(
+                            f"Failed to fetch profile picture from {picture_url}"
+                        )
+                        return "/user.png"
+        except Exception as e:
+            log.error(f"Error processing profile picture '{picture_url}': {e}")
+            return "/user.png"
+
    async def handle_login(self, request, provider):
        if provider not in OAUTH_PROVIDERS:
            raise HTTPException(404)
@@ -257,9 +372,11 @@ class OAuthManager:
                try:
                    access_token = token.get("access_token")
                    headers = {"Authorization": f"Bearer {access_token}"}
-                    async with aiohttp.ClientSession() as session:
+                    async with aiohttp.ClientSession(trust_env=True) as session:
                        async with session.get(
-                            "https://api.github.com/user/emails", headers=headers
+                            "https://api.github.com/user/emails",
+                            headers=headers,
+                            ssl=AIOHTTP_CLIENT_SESSION_SSL,
                        ) as resp:
                            if resp.ok:
                                emails = await resp.json()
@@ -315,6 +432,22 @@ class OAuthManager:
            if user.role != determined_role:
                Users.update_user_role_by_id(user.id, determined_role)

+            # Update profile picture if enabled and different from current
+            if auth_manager_config.OAUTH_UPDATE_PICTURE_ON_LOGIN:
+                picture_claim = auth_manager_config.OAUTH_PICTURE_CLAIM
+                if picture_claim:
+                    new_picture_url = user_data.get(
+                        picture_claim, OAUTH_PROVIDERS[provider].get("picture_url", "")
+                    )
+                    processed_picture_url = await self._process_picture_url(
+                        new_picture_url, token.get("access_token")
+                    )
+                    if processed_picture_url != user.profile_image_url:
+                        Users.update_user_profile_image_url_by_id(
+                            user.id, processed_picture_url
+                        )
+                        log.debug(f"Updated profile picture for user {user.email}")
+
        if not user:
            user_count = Users.get_num_users()

@@ -330,40 +463,9 @@ class OAuthManager:
                    picture_url = user_data.get(
                        picture_claim, OAUTH_PROVIDERS[provider].get("picture_url", "")
                    )
-                    if picture_url:
-                        # Download the profile image into a base64 string
-                        try:
-                            access_token = token.get("access_token")
-                            get_kwargs = {}
-                            if access_token:
-                                get_kwargs["headers"] = {
-                                    "Authorization": f"Bearer {access_token}",
-                                }
-                            async with aiohttp.ClientSession() as session:
-                                async with session.get(
-                                    picture_url, **get_kwargs
-                                ) as resp:
-                                    if resp.ok:
-                                        picture = await resp.read()
-                                        base64_encoded_picture = base64.b64encode(
-                                            picture
-                                        ).decode("utf-8")
-                                        guessed_mime_type = mimetypes.guess_type(
-                                            picture_url
-                                        )[0]
-                                        if guessed_mime_type is None:
-                                            # assume JPG, browsers are tolerant enough of image formats
-                                            guessed_mime_type = "image/jpeg"
-                                        picture_url = f"data:{guessed_mime_type};base64,{base64_encoded_picture}"
-                                    else:
-                                        picture_url = "/user.png"
-                        except Exception as e:
-                            log.error(
-                                f"Error downloading profile image '{picture_url}': {e}"
-                            )
-                            picture_url = "/user.png"
-                    if not picture_url:
-                        picture_url = "/user.png"
+                    picture_url = await self._process_picture_url(
+                        picture_url, token.get("access_token")
+                    )
                else:
                    picture_url = "/user.png"

@@ -59,6 +59,7 @@ def apply_model_params_to_body_openai(params: dict, form_data: dict) -> dict:
        "top_p": float,
        "max_tokens": int,
        "frequency_penalty": float,
+        "presence_penalty": float,
        "reasoning_effort": str,
        "seed": lambda x: x,
        "stop": lambda x: [bytes(s, "utf-8").decode("unicode_escape") for s in x],
@@ -157,7 +157,8 @@ def load_function_module_by_id(function_id, content=None):
            raise Exception("No Function class found in the module")
    except Exception as e:
        log.error(f"Error loading module: {function_id}: {e}")
-        del sys.modules[module_name]  # Cleanup by removing the module in case of error
+        # Cleanup by removing the module in case of error
+        del sys.modules[module_name]

        Functions.update_function_by_id(function_id, {"is_active": False})
        raise e
@@ -182,3 +183,32 @@ def install_frontmatter_requirements(requirements: str):

    else:
        log.info("No requirements found in frontmatter.")
+
+
+def install_tool_and_function_dependencies():
+    """
+    Install all dependencies for all admin tools and active functions.
+
+    By first collecting all dependencies from the frontmatter of each tool and function,
+    and then installing them using pip. Duplicates or similar version specifications are
+    handled by pip as much as possible.
+    """
+    function_list = Functions.get_functions(active_only=True)
+    tool_list = Tools.get_tools()
+
+    all_dependencies = ""
+    try:
+        for function in function_list:
+            frontmatter = extract_frontmatter(replace_imports(function.content))
+            if dependencies := frontmatter.get("requirements"):
+                all_dependencies += f"{dependencies}, "
+        for tool in tool_list:
+            # Only install requirements for admin tools
+            if tool.user.role == "admin":
+                frontmatter = extract_frontmatter(replace_imports(tool.content))
+                if dependencies := frontmatter.get("requirements"):
+                    all_dependencies += f"{dependencies}, "
+
+        install_frontmatter_requirements(all_dependencies.strip(", "))
+    except Exception as e:
+        log.error(f"Error installing requirements: {e}")
@@ -36,11 +36,16 @@ from langchain_core.utils.function_calling import (
 from open_webui.models.tools import Tools
 from open_webui.models.users import UserModel
 from open_webui.utils.plugin import load_tool_module_by_id
-from open_webui.env import AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER_DATA
+from open_webui.env import (
+    SRC_LOG_LEVELS,
+    AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER_DATA,
+    AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL,
+)

 import copy

 log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["MODELS"])


 def get_async_tool_function_and_apply_extra_params(
@@ -276,8 +281,8 @@ def convert_function_to_pydantic_model(func: Callable) -> type[BaseModel]:

    docstring = func.__doc__

-    description = parse_description(docstring)
-    function_descriptions = parse_docstring(docstring)
+    function_description = parse_description(docstring)
+    function_param_descriptions = parse_docstring(docstring)

    field_defs = {}
    for name, param in parameters.items():
@@ -285,15 +290,17 @@ def convert_function_to_pydantic_model(func: Callable) -> type[BaseModel]:
        type_hint = type_hints.get(name, Any)
        default_value = param.default if param.default is not param.empty else ...

-        description = function_descriptions.get(name, None)
+        param_description = function_param_descriptions.get(name, None)

-        if description:
-            field_defs[name] = type_hint, Field(default_value, description=description)
+        if param_description:
+            field_defs[name] = type_hint, Field(
+                default_value, description=param_description
+            )
        else:
            field_defs[name] = type_hint, default_value

    model = create_model(func.__name__, **field_defs)
-    model.__doc__ = description
+    model.__doc__ = function_description

    return model

@@ -371,51 +378,64 @@ def convert_openapi_to_tool_payload(openapi_spec):

    for path, methods in openapi_spec.get("paths", {}).items():
        for method, operation in methods.items():
-            tool = {
-                "type": "function",
-                "name": operation.get("operationId"),
-                "description": operation.get(
-                    "description", operation.get("summary", "No description available.")
-                ),
-                "parameters": {"type": "object", "properties": {}, "required": []},
-            }
-
-            # Extract path and query parameters
-            for param in operation.get("parameters", []):
-                param_name = param["name"]
-                param_schema = param.get("schema", {})
-                tool["parameters"]["properties"][param_name] = {
-                    "type": param_schema.get("type"),
-                    "description": param_schema.get("description", ""),
+            if operation.get("operationId"):
+                tool = {
+                    "type": "function",
+                    "name": operation.get("operationId"),
+                    "description": operation.get(
+                        "description",
+                        operation.get("summary", "No description available."),
+                    ),
+                    "parameters": {"type": "object", "properties": {}, "required": []},
                }
-                if param.get("required"):
-                    tool["parameters"]["required"].append(param_name)

-            # Extract and resolve requestBody if available
-            request_body = operation.get("requestBody")
-            if request_body:
-                content = request_body.get("content", {})
-                json_schema = content.get("application/json", {}).get("schema")
-                if json_schema:
-                    resolved_schema = resolve_schema(
-                        json_schema, openapi_spec.get("components", {})
-                    )
-
-                    if resolved_schema.get("properties"):
-                        tool["parameters"]["properties"].update(
-                            resolved_schema["properties"]
+                # Extract path and query parameters
+                for param in operation.get("parameters", []):
+                    param_name = param["name"]
+                    param_schema = param.get("schema", {})
+                    description = param_schema.get("description", "")
+                    if not description:
+                        description = param.get("description") or ""
+                    if param_schema.get("enum") and isinstance(
+                        param_schema.get("enum"), list
+                    ):
+                        description += (
+                            f". Possible values: {', '.join(param_schema.get('enum'))}"
                        )
-                        if "required" in resolved_schema:
-                            tool["parameters"]["required"] = list(
-                                set(
-                                    tool["parameters"]["required"]
-                                    + resolved_schema["required"]
-                                )
-                            )
-                    elif resolved_schema.get("type") == "array":
-                        tool["parameters"] = resolved_schema  # special case for array
+                    tool["parameters"]["properties"][param_name] = {
+                        "type": param_schema.get("type"),
+                        "description": description,
+                    }
+                    if param.get("required"):
+                        tool["parameters"]["required"].append(param_name)

-            tool_payload.append(tool)
+                # Extract and resolve requestBody if available
+                request_body = operation.get("requestBody")
+                if request_body:
+                    content = request_body.get("content", {})
+                    json_schema = content.get("application/json", {}).get("schema")
+                    if json_schema:
+                        resolved_schema = resolve_schema(
+                            json_schema, openapi_spec.get("components", {})
+                        )
+
+                        if resolved_schema.get("properties"):
+                            tool["parameters"]["properties"].update(
+                                resolved_schema["properties"]
+                            )
+                            if "required" in resolved_schema:
+                                tool["parameters"]["required"] = list(
+                                    set(
+                                        tool["parameters"]["required"]
+                                        + resolved_schema["required"]
+                                    )
+                                )
+                        elif resolved_schema.get("type") == "array":
+                            tool["parameters"] = (
+                                resolved_schema  # special case for array
+                            )
+
+                tool_payload.append(tool)

    return tool_payload

@@ -431,8 +451,10 @@ async def get_tool_server_data(token: str, url: str) -> Dict[str, Any]:
    error = None
    try:
        timeout = aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER_DATA)
-        async with aiohttp.ClientSession(timeout=timeout) as session:
-            async with session.get(url, headers=headers) as response:
+        async with aiohttp.ClientSession(timeout=timeout, trust_env=True) as session:
+            async with session.get(
+                url, headers=headers, ssl=AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL
+            ) as response:
                if response.status != 200:
                    error_body = await response.json()
                    raise Exception(error_body)
@@ -457,7 +479,7 @@ async def get_tool_server_data(token: str, url: str) -> Dict[str, Any]:
        "specs": convert_openapi_to_tool_payload(res),
    }

-    print("Fetched data:", data)
+    log.info("Fetched data:", data)
    return data


@@ -490,7 +512,7 @@ async def get_tool_servers_data(
    results = []
    for (idx, server, url, _), response in zip(server_entries, responses):
        if isinstance(response, Exception):
-            print(f"Failed to connect to {url} OpenAPI tool server")
+            log.error(f"Failed to connect to {url} OpenAPI tool server")
            continue

        results.append(
@@ -573,19 +595,26 @@ async def execute_tool_server(
        if token:
            headers["Authorization"] = f"Bearer {token}"

-        async with aiohttp.ClientSession() as session:
+        async with aiohttp.ClientSession(trust_env=True) as session:
            request_method = getattr(session, http_method.lower())

            if http_method in ["post", "put", "patch"]:
                async with request_method(
-                    final_url, json=body_params, headers=headers
+                    final_url,
+                    json=body_params,
+                    headers=headers,
+                    ssl=AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL,
                ) as response:
                    if response.status >= 400:
                        text = await response.text()
                        raise Exception(f"HTTP error {response.status}: {text}")
                    return await response.json()
            else:
-                async with request_method(final_url, headers=headers) as response:
+                async with request_method(
+                    final_url,
+                    headers=headers,
+                    ssl=AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL,
+                ) as response:
                    if response.status >= 400:
                        text = await response.text()
                        raise Exception(f"HTTP error {response.status}: {text}")
@@ -593,5 +622,5 @@ async def execute_tool_server(

    except Exception as err:
        error = str(err)
-        print("API Request Error:", error)
+        log.exception("API Request Error:", error)
        return {"error": error}
@@ -31,17 +31,18 @@ APScheduler==3.10.4

 RestrictedPython==8.0

-loguru==0.7.2
+loguru==0.7.3
 asgiref==3.8.1

 # AI libraries
 openai
 anthropic
-google-generativeai==0.8.4
+google-genai==1.15.0
+google-generativeai==0.8.5
 tiktoken

-langchain==0.3.19
-langchain-community==0.3.18
+langchain==0.3.24
+langchain-community==0.3.23

 fake-useragent==2.1.0
 chromadb==0.6.3
@@ -49,11 +50,11 @@ pymilvus==2.5.0
 qdrant-client~=1.12.0
 opensearch-py==2.8.0
 playwright==1.49.1 # Caution: version must match docker-compose.playwright.yaml
-elasticsearch==8.17.1
-
+elasticsearch==9.0.1
+pinecone==6.0.2

 transformers
-sentence-transformers==3.3.1
+sentence-transformers==4.1.0
 accelerate
 colbert-ai==0.2.21
 einops==0.8.1
@@ -81,7 +82,7 @@ azure-ai-documentintelligence==1.0.0

 pillow==11.1.0
 opencv-python-headless==4.11.0.86
-rapidocr-onnxruntime==1.3.24
+rapidocr-onnxruntime==1.4.4
 rank-bm25==0.2.2

 onnxruntime==1.20.1
@@ -98,7 +99,7 @@ pytube==15.0.0

 extract_msg
 pydub
-duckduckgo-search~=8.0.0
+duckduckgo-search==8.0.2

 ## Google Drive
 google-api-python-client
@@ -107,7 +108,7 @@ google-auth-oauthlib

 ## Tests
 docker~=7.1.0
-pytest~=8.3.2
+pytest~=8.3.5
 pytest-docker~=3.1.1

 googleapis-common-protos==1.63.2
@@ -127,14 +128,14 @@ firecrawl-py==1.12.0
 tencentcloud-sdk-python==3.0.1336

 ## Trace
-opentelemetry-api==1.31.1
-opentelemetry-sdk==1.31.1
-opentelemetry-exporter-otlp==1.31.1
-opentelemetry-instrumentation==0.52b1
-opentelemetry-instrumentation-fastapi==0.52b1
-opentelemetry-instrumentation-sqlalchemy==0.52b1
-opentelemetry-instrumentation-redis==0.52b1
-opentelemetry-instrumentation-requests==0.52b1
-opentelemetry-instrumentation-logging==0.52b1
-opentelemetry-instrumentation-httpx==0.52b1
-opentelemetry-instrumentation-aiohttp-client==0.52b1
+opentelemetry-api==1.32.1
+opentelemetry-sdk==1.32.1
+opentelemetry-exporter-otlp==1.32.1
+opentelemetry-instrumentation==0.53b1
+opentelemetry-instrumentation-fastapi==0.53b1
+opentelemetry-instrumentation-sqlalchemy==0.53b1
+opentelemetry-instrumentation-redis==0.53b1
+opentelemetry-instrumentation-requests==0.53b1
+opentelemetry-instrumentation-logging==0.53b1
+opentelemetry-instrumentation-httpx==0.53b1
+opentelemetry-instrumentation-aiohttp-client==0.53b1
@@ -65,4 +65,6 @@ if [ -n "$SPACE_ID" ]; then
  export WEBUI_URL=${SPACE_HOST}
 fi

-WEBUI_SECRET_KEY="$WEBUI_SECRET_KEY" exec uvicorn open_webui.main:app --host "$HOST" --port "$PORT" --forwarded-allow-ips '*' --workers "${UVICORN_WORKERS:-1}"
+PYTHON_CMD=$(command -v python3 || command -v python)
+
+WEBUI_SECRET_KEY="$WEBUI_SECRET_KEY" exec "$PYTHON_CMD" -m uvicorn open_webui.main:app --host "$HOST" --port "$PORT" --forwarded-allow-ips '*' --workers "${UVICORN_WORKERS:-1}"
@@ -0,0 +1,74 @@
+import os
+import subprocess
+from collections import Counter
+
+CONFIG_FILE_EXTENSIONS = (".json", ".yml", ".yaml", ".ini", ".conf", ".toml")
+
+
+def is_text_file(filepath):
+    # Check for binary file by scanning for null bytes.
+    try:
+        with open(filepath, "rb") as f:
+            chunk = f.read(4096)
+        if b"\0" in chunk:
+            return False
+        return True
+    except Exception:
+        return False
+
+
+def should_skip_file(path):
+    base = os.path.basename(path)
+    # Skip dotfiles and dotdirs
+    if base.startswith("."):
+        return True
+    # Skip config files by extension
+    if base.lower().endswith(CONFIG_FILE_EXTENSIONS):
+        return True
+    return False
+
+
+def get_tracked_files():
+    try:
+        output = subprocess.check_output(["git", "ls-files"], text=True)
+        files = output.strip().split("\n")
+        files = [f for f in files if f and os.path.isfile(f)]
+        return files
+    except subprocess.CalledProcessError:
+        print("Error: Are you in a git repository?")
+        return []
+
+
+def main():
+    files = get_tracked_files()
+    email_counter = Counter()
+    total_lines = 0
+
+    for file in files:
+        if should_skip_file(file):
+            continue
+        if not is_text_file(file):
+            continue
+        try:
+            blame = subprocess.check_output(
+                ["git", "blame", "-e", file], text=True, errors="replace"
+            )
+            for line in blame.splitlines():
+                # The email always inside <>
+                if "<" in line and ">" in line:
+                    try:
+                        email = line.split("<")[1].split(">")[0].strip()
+                    except Exception:
+                        continue
+                    email_counter[email] += 1
+                    total_lines += 1
+        except subprocess.CalledProcessError:
+            continue
+
+    for email, lines in email_counter.most_common():
+        percent = (lines / total_lines * 100) if total_lines else 0
+        print(f"{email}: {lines}/{total_lines} {percent:.2f}%")
+
+
+if __name__ == "__main__":
+    main()
@@ -1,6 +1,6 @@
 # Hosting UI and Models separately

-Sometimes, its beneficial to host Ollama, separate from the UI, but retain the RAG and RBAC support features shared across users:
+Sometimes, it's beneficial to host Ollama, separate from the UI, but retain the RAG and RBAC support features shared across users:

 # Open WebUI Configuration

@@ -1,11 +1,12 @@
 {
 	"name": "open-webui",
-	"version": "0.6.4",
+	"version": "0.6.10",
 	"private": true,
 	"scripts": {
 		"dev": "npm run pyodide:fetch && vite dev --host",
 		"dev:5050": "npm run pyodide:fetch && vite dev --port 5050",
 		"build": "npm run pyodide:fetch && vite build",
+		"build:watch": "npm run pyodide:fetch && vite build --watch",
 		"preview": "vite preview",
 		"check": "svelte-kit sync && svelte-check --tsconfig ./tsconfig.json",
 		"check:watch": "svelte-kit sync && svelte-check --tsconfig ./tsconfig.json --watch",
@@ -28,8 +29,8 @@
 		"@tailwindcss/container-queries": "^0.1.1",
 		"@tailwindcss/postcss": "^4.0.0",
 		"@tailwindcss/typography": "^0.5.13",
-		"@typescript-eslint/eslint-plugin": "^6.17.0",
-		"@typescript-eslint/parser": "^6.17.0",
+		"@typescript-eslint/eslint-plugin": "^8.31.1",
+		"@typescript-eslint/parser": "^8.31.1",
 		"cypress": "^13.15.0",
 		"eslint": "^8.56.0",
 		"eslint-config-prettier": "^9.1.0",
@@ -61,8 +62,8 @@
 		"@pyscript/core": "^0.4.32",
 		"@sveltejs/adapter-node": "^2.0.0",
 		"@sveltejs/svelte-virtual-list": "^3.0.1",
-		"@tiptap/core": "^2.10.0",
-		"@tiptap/extension-code-block-lowlight": "^2.10.0",
+		"@tiptap/core": "^2.11.9",
+		"@tiptap/extension-code-block-lowlight": "^2.11.9",
 		"@tiptap/extension-highlight": "^2.10.0",
 		"@tiptap/extension-placeholder": "^2.10.0",
 		"@tiptap/extension-typography": "^2.10.0",
@@ -72,13 +73,14 @@
 		"async": "^3.2.5",
 		"bits-ui": "^0.19.7",
 		"codemirror": "^6.0.1",
-		"codemirror-lang-hcl": "^0.0.0-beta.2",
 		"codemirror-lang-elixir": "^4.0.0",
+		"codemirror-lang-hcl": "^0.1.0",
 		"crc-32": "^1.2.2",
 		"dayjs": "^1.11.10",
-		"dompurify": "^3.1.6",
+		"dompurify": "^3.2.5",
 		"eventsource-parser": "^1.1.2",
 		"file-saver": "^2.0.5",
+		"focus-trap": "^7.6.4",
 		"fuse.js": "^7.0.0",
 		"highlight.js": "^11.9.0",
 		"html-entities": "^2.5.3",
@@ -92,7 +94,7 @@
 		"katex": "^0.16.21",
 		"kokoro-js": "^1.1.1",
 		"marked": "^9.1.0",
-		"mermaid": "^10.9.3",
+		"mermaid": "^11.6.0",
 		"paneforge": "^0.0.6",
 		"panzoom": "^9.4.3",
 		"prosemirror-commands": "^1.6.0",
@@ -102,7 +104,7 @@
 		"prosemirror-markdown": "^1.13.1",
 		"prosemirror-model": "^1.23.0",
 		"prosemirror-schema-basic": "^1.2.3",
-		"prosemirror-schema-list": "^1.4.1",
+		"prosemirror-schema-list": "^1.5.1",
 		"prosemirror-state": "^1.4.3",
 		"prosemirror-view": "^1.34.3",
 		"pyodide": "^0.27.3",
@@ -40,16 +40,17 @@ dependencies = [

    "RestrictedPython==8.0",

-    "loguru==0.7.2",
+    "loguru==0.7.3",
    "asgiref==3.8.1",

    "openai",
    "anthropic",
-    "google-generativeai==0.8.4",
+    "google-genai==1.15.0",
+    "google-generativeai==0.8.5",
    "tiktoken",

-    "langchain==0.3.19",
-    "langchain-community==0.3.18",
+    "langchain==0.3.24",
+    "langchain-community==0.3.23",

    "fake-useragent==2.1.0",
    "chromadb==0.6.3",
@@ -57,10 +58,11 @@ dependencies = [
    "qdrant-client~=1.12.0",
    "opensearch-py==2.8.0",
    "playwright==1.49.1",
-    "elasticsearch==8.17.1",
+    "elasticsearch==9.0.1",
+    "pinecone==6.0.2",

    "transformers",
-    "sentence-transformers==3.3.1",
+    "sentence-transformers==4.1.0",
    "accelerate",
    "colbert-ai==0.2.21",
    "einops==0.8.1",
@@ -87,7 +89,7 @@ dependencies = [

    "pillow==11.1.0",
    "opencv-python-headless==4.11.0.86",
-    "rapidocr-onnxruntime==1.3.24",
+    "rapidocr-onnxruntime==1.4.4",
    "rank-bm25==0.2.2",

    "onnxruntime==1.20.1",
@@ -104,7 +106,7 @@ dependencies = [

    "extract_msg",
    "pydub",
-    "duckduckgo-search~=8.0.0",
+    "duckduckgo-search==8.0.2",

    "google-api-python-client",
    "google-auth-httplib2",
@@ -24,6 +24,12 @@
 	font-display: swap;
 }

+@font-face {
+	font-family: 'Vazirmatn';
+	src: url('/assets/fonts/Vazirmatn-Variable.ttf');
+	font-display: swap;
+}
+
 html {
 	word-break: break-word;
 }
@@ -75,7 +81,7 @@ textarea::placeholder {
 }

 .font-primary {
-	font-family: 'Archivo', sans-serif;
+	font-family: 'Archivo', 'Vazirmatn', sans-serif;
 }

 .drag-region {
@@ -308,12 +314,20 @@ input[type='number'] {
 .ProseMirror p.is-editor-empty:first-child::before {
 	content: attr(data-placeholder);
 	float: left;
-	color: #adb5bd;
+	/* Below color is from tailwind, and has the proper contrast
+	text-gray-600 from: https://tailwindcss.com/docs/color */
+	color: #676767;
 	pointer-events: none;

 	@apply line-clamp-1 absolute;
 }

+@media (prefers-color-scheme: dark) {
+	.ProseMirror p.is-editor-empty:first-child::before {
+		color: #757575;
+	}
+}
+
 .ai-autocompletion::after {
 	color: #a0a0a0;

@@ -15,7 +15,7 @@ export const getAudioConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -52,7 +52,7 @@ export const updateAudioConfig = async (token: string, payload: OpenAIConfigForm
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -83,7 +83,7 @@ export const transcribeAudio = async (token: string, file: File) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -120,7 +120,7 @@ export const synthesizeOpenAISpeech = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);

 			return null;
 		});
@@ -152,7 +152,7 @@ export const getModels = async (token: string = ''): Promise<AvailableModelsResp
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);

 			return null;
 		});
@@ -180,7 +180,7 @@ export const getVoices = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);

 			return null;
 		});
@@ -15,7 +15,7 @@ export const getAdminDetails = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -42,7 +42,7 @@ export const getAdminConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -70,7 +70,7 @@ export const updateAdminConfig = async (token: string, body: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -98,7 +98,7 @@ export const getSessionUser = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -129,7 +129,7 @@ export const ldapUserSignIn = async (user: string, password: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);

 			error = err.detail;
 			return null;
@@ -157,7 +157,7 @@ export const getLdapConfig = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -187,7 +187,7 @@ export const updateLdapConfig = async (token: string = '', enable_ldap: boolean)
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -214,7 +214,7 @@ export const getLdapServer = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -242,7 +242,7 @@ export const updateLdapServer = async (token: string = '', body: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -273,7 +273,7 @@ export const userSignIn = async (email: string, password: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);

 			error = err.detail;
 			return null;
@@ -312,7 +312,7 @@ export const userSignUp = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -339,7 +339,7 @@ export const userSignOut = async () => {
 			return res;
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -347,6 +347,7 @@ export const userSignOut = async () => {
 	if (error) {
 		throw error;
 	}
+	return res;
 };

 export const addUser = async (
@@ -354,7 +355,8 @@ export const addUser = async (
 	name: string,
 	email: string,
 	password: string,
-	role: string = 'pending'
+	role: string = 'pending',
+	profile_image_url: null | string = null
 ) => {
 	let error = null;

@@ -368,7 +370,8 @@ export const addUser = async (
 			name: name,
 			email: email,
 			password: password,
-			role: role
+			role: role,
+			...(profile_image_url && { profile_image_url: profile_image_url })
 		})
 	})
 		.then(async (res) => {
@@ -376,7 +379,7 @@ export const addUser = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -407,7 +410,7 @@ export const updateUserProfile = async (token: string, name: string, profileImag
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -438,7 +441,7 @@ export const updateUserPassword = async (token: string, password: string, newPas
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -465,7 +468,7 @@ export const getSignUpEnabledStatus = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -492,7 +495,7 @@ export const getDefaultUserRole = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -522,7 +525,7 @@ export const updateDefaultUserRole = async (token: string, role: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -549,7 +552,7 @@ export const toggleSignUpEnabledStatus = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -576,7 +579,7 @@ export const getJWTExpiresDuration = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -606,7 +609,7 @@ export const updateJWTExpiresDuration = async (token: string, duration: string)
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -633,7 +636,7 @@ export const createAPIKey = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -658,7 +661,7 @@ export const getAPIKey = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -683,7 +686,7 @@ export const deleteAPIKey = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -1,5 +1,4 @@
 import { WEBUI_API_BASE_URL } from '$lib/constants';
-import { t } from 'i18next';

 type ChannelForm = {
 	name: string;
@@ -29,7 +28,7 @@ export const createNewChannel = async (token: string = '', channel: ChannelForm)
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -60,7 +59,7 @@ export const getChannels = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -91,7 +90,7 @@ export const getChannelById = async (token: string = '', channel_id: string) =>
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -127,7 +126,7 @@ export const updateChannelById = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -158,7 +157,7 @@ export const deleteChannelById = async (token: string = '', channel_id: string)
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -197,7 +196,7 @@ export const getChannelMessages = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -237,7 +236,7 @@ export const getChannelThreadMessages = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -276,7 +275,7 @@ export const sendMessage = async (token: string = '', channel_id: string, messag
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -316,7 +315,7 @@ export const updateMessage = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -356,7 +355,7 @@ export const addReaction = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -396,7 +395,7 @@ export const removeReaction = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -430,7 +429,7 @@ export const deleteMessage = async (token: string = '', channel_id: string, mess
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -21,7 +21,7 @@ export const createNewChat = async (token: string, chat: object) => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -61,7 +61,7 @@ export const importChat = async (
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -97,7 +97,7 @@ export const getChatList = async (token: string = '', page: number | null = null
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -131,7 +131,7 @@ export const getChatListByUserId = async (token: string = '', userId: string) =>
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -165,7 +165,7 @@ export const getArchivedChatList = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -196,7 +196,7 @@ export const getAllChats = async (token: string) => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -231,7 +231,7 @@ export const getChatListBySearchText = async (token: string, text: string, page:
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -265,7 +265,7 @@ export const getChatsByFolderId = async (token: string, folderId: string) => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -296,7 +296,7 @@ export const getAllArchivedChats = async (token: string) => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -327,7 +327,7 @@ export const getAllUserChats = async (token: string) => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -358,7 +358,7 @@ export const getAllTags = async (token: string) => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -389,7 +389,7 @@ export const getPinnedChatList = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -426,7 +426,7 @@ export const getChatListByTagName = async (token: string = '', tagName: string)
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -461,7 +461,7 @@ export const getChatById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -493,7 +493,7 @@ export const getChatByShareId = async (token: string, share_id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -531,7 +531,7 @@ export const getChatPinnedStatusById = async (token: string, id: string) => {
 				error = err;
 			}

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -569,7 +569,7 @@ export const toggleChatPinnedStatusById = async (token: string, id: string) => {
 				error = err;
 			}

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -610,7 +610,7 @@ export const cloneChatById = async (token: string, id: string, title?: string) =
 				error = err;
 			}

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -648,7 +648,7 @@ export const cloneSharedChatById = async (token: string, id: string) => {
 				error = err;
 			}

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -680,7 +680,7 @@ export const shareChatById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -715,7 +715,7 @@ export const updateChatFolderIdById = async (token: string, id: string, folderId
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -747,7 +747,7 @@ export const archiveChatById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -779,7 +779,7 @@ export const deleteSharedChatById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -814,7 +814,7 @@ export const updateChatById = async (token: string, id: string, chat: object) =>
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -846,7 +846,7 @@ export const deleteChatById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -878,7 +878,7 @@ export const getTagsById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -912,7 +912,7 @@ export const addTagById = async (token: string, id: string, tagName: string) =>
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -947,7 +947,7 @@ export const deleteTagById = async (token: string, id: string, tagName: string)
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -978,7 +978,7 @@ export const deleteTagsById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -1010,7 +1010,7 @@ export const deleteAllChats = async (token: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -1042,7 +1042,7 @@ export const archiveAllChats = async (token: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -19,7 +19,7 @@ export const importConfig = async (token: string, config) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -46,7 +46,7 @@ export const exportConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -73,7 +73,7 @@ export const getDirectConnectionsConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -103,7 +103,7 @@ export const setDirectConnectionsConfig = async (token: string, config: object)
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -130,7 +130,7 @@ export const getToolServerConnections = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -160,7 +160,7 @@ export const setToolServerConnections = async (token: string, connections: objec
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -190,7 +190,7 @@ export const verifyToolServerConnection = async (token: string, connection: obje
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -217,7 +217,7 @@ export const getCodeExecutionConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -247,7 +247,7 @@ export const setCodeExecutionConfig = async (token: string, config: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -274,7 +274,7 @@ export const getModelsConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -304,7 +304,7 @@ export const setModelsConfig = async (token: string, config: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -334,7 +334,7 @@ export const setDefaultPromptSuggestions = async (token: string, promptSuggestio
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -361,7 +361,7 @@ export const getBanners = async (token: string): Promise<Banner[]> => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -391,7 +391,7 @@ export const setBanners = async (token: string, banners: Banner[]) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -20,7 +20,7 @@ export const getConfig = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -51,7 +51,7 @@ export const updateConfig = async (token: string, config: object) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -82,7 +82,7 @@ export const getAllFeedbacks = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -113,7 +113,7 @@ export const exportAllFeedbacks = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -144,7 +144,7 @@ export const createNewFeedback = async (token: string, feedback: object) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -175,7 +175,7 @@ export const getFeedbackById = async (token: string, feedbackId: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -206,7 +206,7 @@ export const updateFeedbackById = async (token: string, feedbackId: string, feed
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -234,7 +234,7 @@ export const deleteFeedbackById = async (token: string, feedbackId: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -19,7 +19,7 @@ export const uploadFile = async (token: string, file: File) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -76,7 +76,7 @@ export const getFiles = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -107,7 +107,7 @@ export const getFileById = async (token: string, id: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -141,7 +141,7 @@ export const updateFileDataContentById = async (token: string, id: string, conte
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -168,7 +168,7 @@ export const getFileContentById = async (id: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);

 			return null;
 		});
@@ -200,7 +200,7 @@ export const deleteFileById = async (token: string, id: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -231,7 +231,7 @@ export const deleteAllFiles = async (token: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -50,7 +50,7 @@ export const getFolders = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -81,7 +81,7 @@ export const getFolderById = async (token: string, id: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -115,7 +115,7 @@ export const updateFolderNameById = async (token: string, id: string, name: stri
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -153,7 +153,7 @@ export const updateFolderIsExpandedById = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -187,7 +187,7 @@ export const updateFolderParentIdById = async (token: string, id: string, parent
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -226,7 +226,7 @@ export const updateFolderItemsById = async (token: string, id: string, items: Fo
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -257,7 +257,7 @@ export const deleteFolderById = async (token: string, id: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -20,7 +20,7 @@ export const createNewFunction = async (token: string, func: object) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -51,7 +51,7 @@ export const getFunctions = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -82,7 +82,7 @@ export const exportFunctions = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -114,7 +114,7 @@ export const getFunctionById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -149,7 +149,7 @@ export const updateFunctionById = async (token: string, id: string, func: object
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -181,7 +181,7 @@ export const deleteFunctionById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -213,7 +213,7 @@ export const toggleFunctionById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -245,7 +245,7 @@ export const toggleGlobalById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -277,7 +277,7 @@ export const getFunctionValvesById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -309,7 +309,7 @@ export const getFunctionValvesSpecById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -344,7 +344,7 @@ export const updateFunctionValvesById = async (token: string, id: string, valves
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -376,7 +376,7 @@ export const getUserValvesById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -408,7 +408,7 @@ export const getUserValvesSpecById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -443,7 +443,7 @@ export const updateUserValvesById = async (token: string, id: string, valves: ob
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -20,7 +20,7 @@ export const createNewGroup = async (token: string, group: object) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -51,7 +51,7 @@ export const getGroups = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -83,7 +83,7 @@ export const getGroupById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -118,7 +118,7 @@ export const updateGroupById = async (token: string, id: string, group: object)
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -150,7 +150,7 @@ export const deleteGroupById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -16,7 +16,7 @@ export const getConfig = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -51,7 +51,7 @@ export const updateConfig = async (token: string = '', config: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -83,7 +83,7 @@ export const verifyConfigUrl = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -115,7 +115,7 @@ export const getImageGenerationConfig = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -148,7 +148,7 @@ export const updateImageGenerationConfig = async (token: string = '', config: ob
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -180,7 +180,7 @@ export const getImageGenerationModels = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -215,7 +215,7 @@ export const imageGenerations = async (token: string = '', prompt: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -25,7 +25,7 @@ export const getModels = async (
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -173,7 +173,7 @@ export const chatCompleted = async (token: string, body: ChatCompletedForm) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -212,7 +212,7 @@ export const chatAction = async (token: string, action_id: string, body: ChatAct
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -244,7 +244,7 @@ export const stopTask = async (token: string, id: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -276,7 +276,7 @@ export const getTaskIdsByChatId = async (token: string, chat_id: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -315,7 +315,7 @@ export const getToolServerData = async (token: string, url: string) => {
 			}
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -491,7 +491,7 @@ export const getTaskConfig = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -520,7 +520,7 @@ export const updateTaskConfig = async (token: string, config: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -539,7 +539,7 @@ export const updateTaskConfig = async (token: string, config: object) => {
 export const generateTitle = async (
 	token: string = '',
 	model: string,
-	messages: string[],
+	messages: object[],
 	chat_id?: string
 ) => {
 	let error = null;
@@ -562,7 +562,7 @@ export const generateTitle = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			}
@@ -573,7 +573,39 @@ export const generateTitle = async (
 		throw error;
 	}

-	return res?.choices[0]?.message?.content.replace(/["']/g, '') ?? 'New Chat';
+	try {
+		// Step 1: Safely extract the response string
+		const response = res?.choices[0]?.message?.content ?? '';
+
+		// Step 2: Attempt to fix common JSON format issues like single quotes
+		const sanitizedResponse = response.replace(/['‘’`]/g, '"'); // Convert single quotes to double quotes for valid JSON
+
+		// Step 3: Find the relevant JSON block within the response
+		const jsonStartIndex = sanitizedResponse.indexOf('{');
+		const jsonEndIndex = sanitizedResponse.lastIndexOf('}');
+
+		// Step 4: Check if we found a valid JSON block (with both `{` and `}`)
+		if (jsonStartIndex !== -1 && jsonEndIndex !== -1) {
+			const jsonResponse = sanitizedResponse.substring(jsonStartIndex, jsonEndIndex + 1);
+
+			// Step 5: Parse the JSON block
+			const parsed = JSON.parse(jsonResponse);
+
+			// Step 6: If there's a "tags" key, return the tags array; otherwise, return an empty array
+			if (parsed && parsed.title) {
+				return parsed.title;
+			} else {
+				return null;
+			}
+		}
+
+		// If no valid JSON block found, return an empty array
+		return null;
+	} catch (e) {
+		// Catch and safely return empty array on any parsing errors
+		console.error('Failed to parse response: ', e);
+		return null;
+	}
 };

 export const generateTags = async (
@@ -602,7 +634,7 @@ export const generateTags = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			}
@@ -674,7 +706,7 @@ export const generateEmoji = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			}
@@ -724,7 +756,7 @@ export const generateQueries = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			}
@@ -796,7 +828,7 @@ export const generateAutoCompletion = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			}
@@ -860,7 +892,7 @@ export const generateMoACompletion = async (
 			stream: true
 		})
 	}).catch((err) => {
-		console.log(err);
+		console.error(err);
 		error = err;
 		return null;
 	});
@@ -888,7 +920,7 @@ export const getPipelinesList = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -922,7 +954,7 @@ export const uploadPipeline = async (token: string, file: File, urlIdx: string)
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -958,7 +990,7 @@ export const downloadPipeline = async (token: string, url: string, urlIdx: strin
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -994,7 +1026,7 @@ export const deletePipeline = async (token: string, id: string, urlIdx: string)
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -1031,7 +1063,7 @@ export const getPipelines = async (token: string, urlIdx?: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1068,7 +1100,7 @@ export const getPipelineValves = async (token: string, pipeline_id: string, urlI
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1104,7 +1136,7 @@ export const getPipelineValvesSpec = async (token: string, pipeline_id: string,
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1146,7 +1178,7 @@ export const updatePipelineValves = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);

 			if ('detail' in err) {
 				error = err.detail;
@@ -1178,7 +1210,7 @@ export const getBackendConfig = async () => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1204,7 +1236,7 @@ export const getChangelog = async () => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1231,7 +1263,7 @@ export const getVersionUpdates = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1258,7 +1290,7 @@ export const getModelFilterConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1293,7 +1325,7 @@ export const updateModelFilterConfig = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1320,7 +1352,7 @@ export const getWebhookUrl = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1350,7 +1382,7 @@ export const updateWebhookUrl = async (token: string, url: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1377,7 +1409,7 @@ export const getCommunitySharingEnabledStatus = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1404,7 +1436,7 @@ export const toggleCommunitySharingEnabledStatus = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -1431,7 +1463,7 @@ export const getModelConfig = async (token: string): Promise<GlobalModelConfig>
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -1479,7 +1511,7 @@ export const updateModelConfig = async (token: string, config: GlobalModelConfig
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -27,7 +27,7 @@ export const createNewKnowledge = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -58,7 +58,7 @@ export const getKnowledgeBases = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -89,7 +89,7 @@ export const getKnowledgeBaseList = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -121,7 +121,7 @@ export const getKnowledgeById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -166,7 +166,7 @@ export const updateKnowledgeById = async (token: string, id: string, form: Knowl
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -201,7 +201,7 @@ export const addFileToKnowledgeById = async (token: string, id: string, fileId:
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -236,7 +236,7 @@ export const updateFileFromKnowledgeById = async (token: string, id: string, fil
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -271,7 +271,7 @@ export const removeFileFromKnowledgeById = async (token: string, id: string, fil
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -303,7 +303,7 @@ export const resetKnowledgeById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -335,7 +335,7 @@ export const deleteKnowledgeById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -363,7 +363,7 @@ export const reindexKnowledgeFiles = async (token: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -17,7 +17,7 @@ export const getMemories = async (token: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -48,7 +48,7 @@ export const addNewMemory = async (token: string, content: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -79,7 +79,7 @@ export const updateMemoryById = async (token: string, id: string, content: strin
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -110,7 +110,7 @@ export const queryMemory = async (token: string, content: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -142,7 +142,7 @@ export const deleteMemoryById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -174,7 +174,7 @@ export const deleteMemoriesByUserId = async (token: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -20,7 +20,7 @@ export const getModels = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -51,7 +51,7 @@ export const getBaseModels = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -80,7 +80,7 @@ export const createNewModel = async (token: string, model: object) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -115,7 +115,7 @@ export const getModelById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -150,7 +150,7 @@ export const toggleModelById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -186,7 +186,7 @@ export const updateModelById = async (token: string, id: string, model: object)
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -221,7 +221,7 @@ export const deleteModelById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -253,7 +253,7 @@ export const deleteAllModels = async (token: string) => {
 		.catch((err) => {
 			error = err;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -0,0 +1,187 @@
+import { WEBUI_API_BASE_URL } from '$lib/constants';
+import { getTimeRange } from '$lib/utils';
+
+type NoteItem = {
+	title: string;
+	data: object;
+	meta?: null | object;
+	access_control?: null | object;
+};
+
+export const createNewNote = async (token: string, note: NoteItem) => {
+	let error = null;
+
+	const res = await fetch(`${WEBUI_API_BASE_URL}/notes/create`, {
+		method: 'POST',
+		headers: {
+			Accept: 'application/json',
+			'Content-Type': 'application/json',
+			authorization: `Bearer ${token}`
+		},
+		body: JSON.stringify({
+			...note
+		})
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.catch((err) => {
+			error = err.detail;
+			console.error(err);
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	return res;
+};
+
+export const getNotes = async (token: string = '') => {
+	let error = null;
+
+	const res = await fetch(`${WEBUI_API_BASE_URL}/notes/`, {
+		method: 'GET',
+		headers: {
+			Accept: 'application/json',
+			'Content-Type': 'application/json',
+			authorization: `Bearer ${token}`
+		}
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.then((json) => {
+			return json;
+		})
+		.catch((err) => {
+			error = err.detail;
+			console.error(err);
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	if (!Array.isArray(res)) {
+		return {}; // or throw new Error("Notes response is not an array")
+	}
+
+	// Build the grouped object
+	const grouped: Record<string, any[]> = {};
+	for (const note of res) {
+		const timeRange = getTimeRange(note.updated_at / 1000000000);
+		if (!grouped[timeRange]) {
+			grouped[timeRange] = [];
+		}
+		grouped[timeRange].push({
+			...note,
+			timeRange
+		});
+	}
+
+	return grouped;
+};
+
+export const getNoteById = async (token: string, id: string) => {
+	let error = null;
+
+	const res = await fetch(`${WEBUI_API_BASE_URL}/notes/${id}`, {
+		method: 'GET',
+		headers: {
+			Accept: 'application/json',
+			'Content-Type': 'application/json',
+			authorization: `Bearer ${token}`
+		}
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.then((json) => {
+			return json;
+		})
+		.catch((err) => {
+			error = err.detail;
+
+			console.error(err);
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	return res;
+};
+
+export const updateNoteById = async (token: string, id: string, note: NoteItem) => {
+	let error = null;
+
+	const res = await fetch(`${WEBUI_API_BASE_URL}/notes/${id}/update`, {
+		method: 'POST',
+		headers: {
+			Accept: 'application/json',
+			'Content-Type': 'application/json',
+			authorization: `Bearer ${token}`
+		},
+		body: JSON.stringify({
+			...note
+		})
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.then((json) => {
+			return json;
+		})
+		.catch((err) => {
+			error = err.detail;
+
+			console.error(err);
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	return res;
+};
+
+export const deleteNoteById = async (token: string, id: string) => {
+	let error = null;
+
+	const res = await fetch(`${WEBUI_API_BASE_URL}/notes/${id}/delete`, {
+		method: 'DELETE',
+		headers: {
+			Accept: 'application/json',
+			'Content-Type': 'application/json',
+			authorization: `Bearer ${token}`
+		}
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.then((json) => {
+			return json;
+		})
+		.catch((err) => {
+			error = err.detail;
+
+			console.error(err);
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	return res;
+};
@@ -1,10 +1,6 @@
 import { OLLAMA_API_BASE_URL } from '$lib/constants';

-export const verifyOllamaConnection = async (
-	token: string = '',
-	url: string = '',
-	key: string = ''
-) => {
+export const verifyOllamaConnection = async (token: string = '', connection: dict = {}) => {
 	let error = null;

 	const res = await fetch(`${OLLAMA_API_BASE_URL}/verify`, {
@@ -15,8 +11,7 @@ export const verifyOllamaConnection = async (
 			'Content-Type': 'application/json'
 		},
 		body: JSON.stringify({
-			url,
-			key
+			...connection
 		})
 	})
 		.then(async (res) => {
@@ -51,7 +46,7 @@ export const getOllamaConfig = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -92,7 +87,7 @@ export const updateOllamaConfig = async (token: string = '', config: OllamaConfi
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -124,7 +119,7 @@ export const getOllamaUrls = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -159,7 +154,7 @@ export const updateOllamaUrls = async (token: string = '', urls: string[]) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -191,7 +186,7 @@ export const getOllamaVersion = async (token: string, urlIdx?: number) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -223,7 +218,7 @@ export const getOllamaModels = async (token: string = '', urlIdx: null | number
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -268,7 +263,7 @@ export const generatePrompt = async (token: string = '', model: string, conversa
 			`
 		})
 	}).catch((err) => {
-		console.log(err);
+		console.error(err);
 		if ('detail' in err) {
 			error = err.detail;
 		}
@@ -408,11 +403,11 @@ export const deleteModel = async (token: string, tagName: string, urlIdx: string
 			return res.json();
 		})
 		.then((json) => {
-			console.log(json);
+			console.debug(json);
 			return true;
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;

 			if ('detail' in err) {
@@ -445,7 +440,7 @@ export const pullModel = async (token: string, tagName: string, urlIdx: number |
 			name: tagName
 		})
 	}).catch((err) => {
-		console.log(err);
+		console.error(err);
 		error = err;

 		if ('detail' in err) {
@@ -481,7 +476,7 @@ export const downloadModel = async (
 			})
 		}
 	).catch((err) => {
-		console.log(err);
+		console.error(err);
 		error = err;

 		if ('detail' in err) {
@@ -512,7 +507,7 @@ export const uploadModel = async (token: string, file: File, urlIdx: string | nu
 			body: formData
 		}
 	).catch((err) => {
-		console.log(err);
+		console.error(err);
 		error = err;

 		if ('detail' in err) {
@@ -16,7 +16,7 @@ export const getOpenAIConfig = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -58,7 +58,7 @@ export const updateOpenAIConfig = async (token: string = '', config: OpenAIConfi
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -90,7 +90,7 @@ export const getOpenAIUrls = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -125,7 +125,7 @@ export const updateOpenAIUrls = async (token: string = '', urls: string[]) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -157,7 +157,7 @@ export const getOpenAIKeys = async (token: string = '') => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -192,7 +192,7 @@ export const updateOpenAIKeys = async (token: string = '', keys: string[]) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			if ('detail' in err) {
 				error = err.detail;
 			} else {
@@ -267,10 +267,10 @@ export const getOpenAIModels = async (token: string, urlIdx?: number) => {

 export const verifyOpenAIConnection = async (
 	token: string = '',
-	url: string = 'https://api.openai.com/v1',
-	key: string = '',
+	connection: dict = {},
 	direct: boolean = false
 ) => {
+	const { url, key, config } = connection;
 	if (!url) {
 		throw 'OpenAI: URL is required';
 	}
@@ -309,7 +309,8 @@ export const verifyOpenAIConnection = async (
 			},
 			body: JSON.stringify({
 				url,
-				key
+				key,
+				config
 			})
 		})
 			.then(async (res) => {
@@ -346,7 +347,7 @@ export const chatCompletion = async (
 		},
 		body: JSON.stringify(body)
 	}).catch((err) => {
-		console.log(err);
+		console.error(err);
 		error = err;
 		return null;
 	});
@@ -409,7 +410,7 @@ export const synthesizeOpenAISpeech = async (
 			voice: speaker
 		})
 	}).catch((err) => {
-		console.log(err);
+		console.error(err);
 		error = err;
 		return null;
 	});
@@ -28,7 +28,7 @@ export const createNewPrompt = async (token: string, prompt: PromptItem) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -59,7 +59,7 @@ export const getPrompts = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -90,7 +90,7 @@ export const getPromptList = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -122,7 +122,7 @@ export const getPromptByCommand = async (token: string, command: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -158,7 +158,7 @@ export const updatePromptByCommand = async (token: string, prompt: PromptItem) =
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -192,7 +192,7 @@ export const deletePromptByCommand = async (token: string, command: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -15,7 +15,7 @@ export const getRAGConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -77,7 +77,7 @@ export const updateRAGConfig = async (token: string, payload: RAGConfigForm) =>
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -104,7 +104,7 @@ export const getQuerySettings = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -140,7 +140,7 @@ export const updateQuerySettings = async (token: string, settings: QuerySettings
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -167,7 +167,7 @@ export const getEmbeddingConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -209,7 +209,7 @@ export const updateEmbeddingConfig = async (token: string, payload: EmbeddingMod
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -236,7 +236,7 @@ export const getRerankingConfig = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -270,7 +270,7 @@ export const updateRerankingConfig = async (token: string, payload: RerankingMod
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -313,7 +313,7 @@ export const processFile = async (
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -344,7 +344,7 @@ export const processYoutubeVideo = async (token: string, url: string) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -376,7 +376,7 @@ export const processWeb = async (token: string, collection_name: string, url: st
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -410,7 +410,7 @@ export const processWebSearch = async (
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -20,7 +20,7 @@ export const createNewTool = async (token: string, tool: object) => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -51,7 +51,7 @@ export const getTools = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -82,7 +82,7 @@ export const getToolList = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -113,7 +113,7 @@ export const exportTools = async (token: string = '') => {
 		})
 		.catch((err) => {
 			error = err.detail;
-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -145,7 +145,7 @@ export const getToolById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -180,7 +180,7 @@ export const updateToolById = async (token: string, id: string, tool: object) =>
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -212,7 +212,7 @@ export const deleteToolById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -244,7 +244,7 @@ export const getToolValvesById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -276,7 +276,7 @@ export const getToolValvesSpecById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -311,7 +311,7 @@ export const updateToolValvesById = async (token: string, id: string, valves: ob
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -343,7 +343,7 @@ export const getUserValvesById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -375,7 +375,7 @@ export const getUserValvesSpecById = async (token: string, id: string) => {
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -410,7 +410,7 @@ export const updateUserValvesById = async (token: string, id: string, valves: ob
 		.catch((err) => {
 			error = err.detail;

-			console.log(err);
+			console.error(err);
 			return null;
 		});

@@ -16,7 +16,7 @@ export const getUserGroups = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -43,7 +43,7 @@ export const getUserDefaultPermissions = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -73,7 +73,7 @@ export const updateUserDefaultPermissions = async (token: string, permissions: o
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -104,7 +104,7 @@ export const updateUserRole = async (token: string, id: string, role: string) =>
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -116,10 +116,33 @@ export const updateUserRole = async (token: string, id: string, role: string) =>
 	return res;
 };

-export const getUsers = async (token: string) => {
+export const getUsers = async (
+	token: string,
+	query?: string,
+	orderBy?: string,
+	direction?: string,
+	page = 1
+) => {
 	let error = null;
+	let res = null;

-	const res = await fetch(`${WEBUI_API_BASE_URL}/users/`, {
+	let searchParams = new URLSearchParams();
+
+	searchParams.set('page', `${page}`);
+
+	if (query) {
+		searchParams.set('query', query);
+	}
+
+	if (orderBy) {
+		searchParams.set('order_by', orderBy);
+	}
+
+	if (direction) {
+		searchParams.set('direction', direction);
+	}
+
+	res = await fetch(`${WEBUI_API_BASE_URL}/users/?${searchParams.toString()}`, {
 		method: 'GET',
 		headers: {
 			'Content-Type': 'application/json',
@@ -131,7 +154,7 @@ export const getUsers = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -140,7 +163,35 @@ export const getUsers = async (token: string) => {
 		throw error;
 	}

-	return res ? res : [];
+	return res;
+};
+
+export const getAllUsers = async (token: string) => {
+	let error = null;
+	let res = null;
+
+	res = await fetch(`${WEBUI_API_BASE_URL}/users/all`, {
+		method: 'GET',
+		headers: {
+			'Content-Type': 'application/json',
+			Authorization: `Bearer ${token}`
+		}
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.catch((err) => {
+			console.error(err);
+			error = err.detail;
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	return res;
 };

 export const getUserSettings = async (token: string) => {
@@ -157,7 +208,7 @@ export const getUserSettings = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -187,7 +238,7 @@ export const updateUserSettings = async (token: string, settings: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -214,7 +265,7 @@ export const getUserById = async (token: string, userId: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -240,7 +291,7 @@ export const getUserInfo = async (token: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -270,7 +321,7 @@ export const updateUserInfo = async (token: string, info: object) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -284,7 +335,7 @@ export const updateUserInfo = async (token: string, info: object) => {

 export const getAndUpdateUserLocation = async (token: string) => {
 	const location = await getUserPosition().catch((err) => {
-		console.log(err);
+		console.error(err);
 		return null;
 	});

@@ -292,7 +343,7 @@ export const getAndUpdateUserLocation = async (token: string) => {
 		await updateUserInfo(token, { location: location });
 		return location;
 	} else {
-		console.log('Failed to get user location');
+		console.info('Failed to get user location');
 		return null;
 	}
 };
@@ -312,7 +363,7 @@ export const deleteUserById = async (token: string, userId: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -352,7 +403,7 @@ export const updateUserById = async (token: string, userId: string, user: UserUp
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -15,7 +15,7 @@ export const getGravatarUrl = async (token: string, email: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -41,7 +41,7 @@ export const executeCode = async (token: string, code: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);

 			error = err;
 			if (err.detail) {
@@ -75,7 +75,7 @@ export const formatPythonCode = async (token: string, code: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);

 			error = err;
 			if (err.detail) {
@@ -110,7 +110,7 @@ export const downloadChatAsPDF = async (token: string, title: string, messages:
 			return res.blob();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -136,7 +136,7 @@ export const getHTMLFromMarkdown = async (token: string, md: string) => {
 			return res.json();
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err;
 			return null;
 		});
@@ -170,7 +170,7 @@ export const downloadDatabase = async (token: string) => {
 			window.URL.revokeObjectURL(url);
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -206,7 +206,7 @@ export const downloadLiteLLMConfig = async (token: string) => {
 			window.URL.revokeObjectURL(url);
 		})
 		.catch((err) => {
-			console.log(err);
+			console.error(err);
 			error = err.detail;
 			return null;
 		});
@@ -30,8 +30,17 @@
 	let url = '';
 	let key = '';

+	let connectionType = 'external';
+	let azure = false;
+	$: azure =
+		(url.includes('azure.com') || url.includes('cognitive.microsoft.com')) && !direct
+			? true
+			: false;
+
 	let prefixId = '';
 	let enable = true;
+	let apiVersion = '';
+
 	let tags = [];

 	let modelId = '';
@@ -40,7 +49,10 @@
 	let loading = false;

 	const verifyOllamaHandler = async () => {
-		const res = await verifyOllamaConnection(localStorage.token, url, key).catch((error) => {
+		const res = await verifyOllamaConnection(localStorage.token, {
+			url,
+			key
+		}).catch((error) => {
 			toast.error(`${error}`);
 		});

@@ -50,11 +62,20 @@
 	};

 	const verifyOpenAIHandler = async () => {
-		const res = await verifyOpenAIConnection(localStorage.token, url, key, direct).catch(
-			(error) => {
-				toast.error(`${error}`);
-			}
-		);
+		const res = await verifyOpenAIConnection(
+			localStorage.token,
+			{
+				url,
+				key,
+				config: {
+					azure: azure,
+					api_version: apiVersion
+				}
+			},
+			direct
+		).catch((error) => {
+			toast.error(`${error}`);
+		});

 		if (res) {
 			toast.success($i18n.t('Server connection verified'));
@@ -85,6 +106,28 @@
 			return;
 		}

+		if (azure) {
+			if (!apiVersion) {
+				loading = false;
+
+				toast.error('API Version is required');
+				return;
+			}
+
+			if (!key) {
+				loading = false;
+
+				toast.error('Key is required');
+				return;
+			}
+
+			if (modelIds.length === 0) {
+				loading = false;
+				toast.error('Deployment names are required');
+				return;
+			}
+		}
+
 		// remove trailing slash from url
 		url = url.replace(/\/$/, '');

@@ -95,7 +138,9 @@
 				enable: enable,
 				tags: tags,
 				prefix_id: prefixId,
-				model_ids: modelIds
+				model_ids: modelIds,
+				connection_type: connectionType,
+				...(!ollama && azure ? { azure: true, api_version: apiVersion } : {})
 			}
 		};

@@ -120,6 +165,14 @@
 			tags = connection.config?.tags ?? [];
 			prefixId = connection.config?.prefix_id ?? '';
 			modelIds = connection.config?.model_ids ?? [];
+
+			if (ollama) {
+				connectionType = connection.config?.connection_type ?? 'local';
+			} else {
+				connectionType = connection.config?.connection_type ?? 'external';
+				azure = connection.config?.azure ?? false;
+				apiVersion = connection.config?.api_version ?? '';
+			}
 		}
 	};

@@ -134,7 +187,7 @@

 <Modal size="sm" bind:show>
 	<div>
-		<div class=" flex justify-between dark:text-gray-100 px-5 pt-4 pb-2">
+		<div class=" flex justify-between dark:text-gray-100 px-5 pt-4 pb-1.5">
 			<div class=" text-lg font-medium self-center font-primary">
 				{#if edit}
 					{$i18n.t('Edit Connection')}
@@ -171,7 +224,31 @@
 					}}
 				>
 					<div class="px-1">
-						<div class="flex gap-2">
+						{#if !direct}
+							<div class="flex gap-2">
+								<div class="flex w-full justify-between items-center">
+									<div class=" text-xs text-gray-500">{$i18n.t('Connection Type')}</div>
+
+									<div class="">
+										<button
+											on:click={() => {
+												connectionType = connectionType === 'local' ? 'external' : 'local';
+											}}
+											type="button"
+											class=" text-xs text-gray-700 dark:text-gray-300"
+										>
+											{#if connectionType === 'local'}
+												{$i18n.t('Local')}
+											{:else}
+												{$i18n.t('External')}
+											{/if}
+										</button>
+									</div>
+								</div>
+							</div>
+						{/if}
+
+						<div class="flex gap-2 mt-1.5">
 							<div class="flex flex-col w-full">
 								<div class=" mb-0.5 text-xs text-gray-500">{$i18n.t('URL')}</div>

@@ -252,6 +329,25 @@
 							</div>
 						</div>

+						{#if azure}
+							<div class="flex gap-2 mt-2">
+								<div class="flex flex-col w-full">
+									<div class=" mb-1 text-xs text-gray-500">{$i18n.t('API Version')}</div>
+
+									<div class="flex-1">
+										<input
+											class="w-full text-sm bg-transparent placeholder:text-gray-300 dark:placeholder:text-gray-700 outline-hidden"
+											type="text"
+											bind:value={apiVersion}
+											placeholder={$i18n.t('API Version')}
+											autocomplete="off"
+											required
+										/>
+									</div>
+								</div>
+							</div>
+						{/if}
+
 						<div class="flex gap-2 mt-2">
 							<div class="flex flex-col w-full">
 								<div class=" mb-1.5 text-xs text-gray-500">{$i18n.t('Tags')}</div>
@@ -308,6 +404,11 @@
 										{$i18n.t('Leave empty to include all models from "{{url}}/api/tags" endpoint', {
 											url: url
 										})}
+									{:else if azure}
+										{$i18n.t('Deployment names are required for Azure OpenAI')}
+										<!-- {$i18n.t('Leave empty to include all models from "{{url}}" endpoint', {
+											url: `${url}/openai/deployments`
+										})} -->
 									{:else}
 										{$i18n.t('Leave empty to include all models from "{{url}}/models" endpoint', {
 											url: url
@@ -317,7 +418,7 @@
 							{/if}
 						</div>

-						<hr class=" border-gray-100 dark:border-gray-700/10 my-2.5 w-full" />
+						<hr class=" border-gray-100 dark:border-gray-700/10 my-1.5 w-full" />

 						<div class="flex items-center">
 							<input
@@ -21,7 +21,7 @@
 			{#if content}
 				{content}
 			{:else}
-				{$i18n.t('Drop any files here to add to the conversation')}
+				{$i18n.t('Drop any files here to upload')}
 			{/if}
 		</div>
 	</slot>
--- a/Show More
+++ b/Show More