Releases · owndev/Open-WebUI-Functions

📦 Open-WebUI-Functions v1.0.0

Warning

Parts of these release notes were generated with AI and may contain mistakes or inaccuracies. Please verify important details against the actual commits, pull requests, and source changes.

Open-WebUI-Functions v1.0.0 introduces a robust set of modular Python pipelines, filters, and integrations that significantly extend the capabilities of Open WebUI. This first stable release delivers seamless connectivity to external AI providers—including Azure OpenAI, Google Gemini, Infomaniak AI Tools, and N8N—alongside advanced features like native citation support, multimodal processing, secure secret management, and detailed agent tool usage display. The architecture is designed for flexibility, streaming and non-streaming data flows, and granular authentication, empowering administrators and developers to automate workflows and enhance frontend transparency.

✨ New Features

Added modular pipelines for Azure OpenAI, Google Gemini, Infomaniak AI Tools, and N8N, enabling customizable request/response processing, model inference, and event emission.
Introduced Azure AI Search integration with native citation support, including structured events for frontend citation cards and relevance scoring.
Integrated Google Gemini pipeline with multimodal (image/video) generation, configurable aspect ratio/resolution, and support for Veo models.
Enabled N8N workflow automation with detailed AI agent tool usage display, configurable verbosity, and session handling.
Implemented flexible authentication options, including support for Cloudflare Access and encrypted API key management.
Added streaming support across pipelines, allowing real-time data processing and prompt-feedback handling.
Provided native tool calling and Google search grounding for enhanced RAG (Retrieval Augmented Generation) scenarios.
Supported conversation history inclusion in N8N requests for improved context-aware automation.
Emitted structured events for citation linking, tool usage transparency, and chat lifecycle, facilitating advanced frontend features.

🚀 Improvements

Enhanced Google Gemini pipeline with improved model compatibility, thinking level validation, and configurable thinking budgets.
Refactored video generation event emission to a dedicated function for cleaner code and better error handling.
Improved Azure AI and Gemini pipelines with better model handling, usage data extraction, and optional event emitter integration for Infomaniak API.
Added model whitelist and additional model support to Gemini pipeline for granular control over available models.
Updated password input handling and usage tracking across pipelines for improved security and analytics.
Optimized citation handling in Azure AI pipeline, including normalization, markdown link conversion, and XSS protection via html.escape.
Improved logging efficiency by avoiding unnecessary JSON serialization and providing comprehensive debug logs for citation events.
Enhanced streaming response handling in Infomaniak pipeline with error suppression during cleanup.
Updated API endpoints and versioning for Infomaniak integration to support v2.1.0.
Added support for Enterprise Web Search on Vertex AI, expanding search capabilities.
Improved score matching and normalization logic for Azure AI Search citations, ensuring accurate relevance display.
Refactored system prompt hierarchy and fallback logic in Gemini pipeline for more reliable prompt management.

🔧 Bug Fixes

Corrected badge URLs and removed outdated badges in README.md for accurate project status display.
Fixed image generation model filtering and default value handling in Gemini pipeline.
Addressed incorrect file retrieval from S3 and undefined variable errors in Azure AI pipeline.
Resolved citation structure issues, including event emission per source and proper OpenWebUI display.
Fixed streaming parsing and buffer handling for N8N responses.
Corrected merge issues, excluded private methods, and improved code review compliance.
Fixed thinking config, copilot recommendations, and prompt-feedback streaming in Gemini pipeline.
Resolved NoneType errors and improved handling of null citation fields.
Fixed multi-image upload handling for image generation models.
Streamlined Azure OpenAI endpoint checks and removed unnecessary imports.

🔄 Migration Notes

For users upgrading from custom prototypes or earlier forks, ensure environment variables and pipeline configurations are updated to reflect new authentication options and event emitter logic.

Citation event structure has changed: now emits separate events per source with unique names and markdown links. Update any custom frontend handlers to match the new format:

# Example: Citation event structure
event = {
    "type": "citation",
    "source": "[doc1]",
    "url": "https://example.com/doc1",
    "distances": [0.85, 0.92],
    ...
}

Gemini pipeline system prompt logic now prioritizes chat controls, model defaults, and user overrides. If you previously relied on environment variables for prompt injection, review the new hierarchy and adjust accordingly.

⚠️ Breaking Changes

Removed deprecated valves (e.g., AZURE_AI_ENHANCE_CITATIONS, AZURE_AI_OPENWEBUI_CITATIONS, AZURE_AI_LINK_CITATIONS) and simplified citation handling; update any scripts referencing these valves.
Changed Gemini pipeline configuration: replaced THINKING_ENABLED with INCLUDE_THOUGHTS for thought output. Update environment variables to match:
```
# Old
THINKING_ENABLED=true
# New
INCLUDE_THOUGHTS=true
```
Citation event format and linking have been overhauled; custom integrations must adapt to markdown link structure and new event emission patterns.

📦 Dependencies

Upgraded Google Gemini pipeline to version 1.11.1, Infomaniak API interface to v2.1.0, and Azure AI Foundry pipeline to latest supported version.
Switched to the newer python-genai SDK for Google Gemini, enabling Vertex AI support.
Added minimum Open WebUI version requirements across multiple pipelines for compatibility.

🔒 Security

Implemented automatic encryption for API keys and sensitive secrets, ensuring secure storage and transmission.
Enhanced XSS protection in citation tag names via html escaping.
Added support for Cloudflare Access authentication, improving access control for administrator endpoints.
Replaced deprecated decryption methods and improved password input handling.

🏗️ Infrastructure & CI/CD

Added GitHub workflows for automated releases and versioning, streamlining deployment and release management.
Updated badge URLs and issue templates for improved project documentation and contributor guidance.

📝 Other Changes

Revised README.md for clarity, structure, and updated feature descriptions.
Added detailed integration documentation, including index schema and field mapping for Azure AI citations.
Enhanced code comments, type annotations, and efficiency improvements per code review feedback.
Removed dead code and unused functions for maintainability.
Added technical spike documentation and improved Copilot prompts.

📋 Release Information

Application: Open-WebUI-Functions
Version: 1.0.0
Repository: owndev/Open-WebUI-Functions
Commit: a2154885d69f206b9b9d412e0e9cb59f7f3f6985
Release Type: ✅ Stable Release
Environment: Production
Created: 2026-03-16T08:33:54+01:00

📋 Version Information

Full Version: 1.0.0
Assembly Version: 1.0.0.60000
Informational Version: 1.0.0+Branch.main.Sha.a2154885d69f206b9b9d412e0e9cb59f7f3f6985
Pre-release Tag: None
Pre-release Number: 0
Commits Since Version: 0
Weighted Pre-release Number: 60000

🛠️ Build Information

Workflow: 🏷️ GitHub Release
Run ID: 23132851093
Actor: @owndev

🐛 Support & Feedback

📋 Need Help?	🔗 Resource
🐛 Report Issues	owndev/Open-WebUI-Functions/issues
💬 Discussions	owndev/Open-WebUI-Functions/discussions
📚 Documentation	owndev/Open-WebUI-Functions/wiki

What's Changed

[FIX] Support Including Model Name in Request Body for Single-Deployment Endpoints by @guoanhu12146 in #3
Add all azure models as pipes() by @LionsAd in #5
Remove azure_ai_foundry_deepseek.py (no longer needed) by @Phlogi in #9
Add new function for Infomaniak AI Tools by @owndev in #8
Add encryption for API Keys by @owndev in #11
Add new Models by @ssweber in #13
18 feature request multiple model support by @owndev in #19
21 feature request reasoning effort and chain of thought with o4 mini by @owndev in #22
20 time tracker with log analytics azure by @owndev in #23
Add event_emitter to Azure AI Foundry Pipeline by @ssweber in #29
Add d...