Open-WebUI-Functions provides a suite of custom Python pipelines, filters, and integrations to significantly enhance Open WebUI. It enables seamless interaction with various AI models like Azure AI and Google Gemini, and automation platforms such as N8N, offering dynamic request handling, data preprocessing, and workflow automation for advanced users.

How It Works

This project injects custom Python functions into Open WebUI via its admin interface, acting as extensions for AI processing and data manipulation. Core features include configurable pipelines for model inference and data transformations, filters for preprocessing/postprocessing, and direct integrations with services like Azure OpenAI (with RAG), N8N, and Google Gemini. It supports both streaming and non-streaming responses and includes automatic API key encryption when a secret key is configured.

Quick Start & Requirements

• Installation: Requires Open WebUI admin access. Navigate to Admin Settings -> Function Tab, click "Add New Function," and paste the function code.
• Prerequisites: An active Open WebUI instance, admin privileges, and potentially external AI services (e.g., Azure AI).
• Configuration: Set WEBUI_SECRET_KEY environment variable for encryption. Additional environment variables may be needed per pipeline.
• Links:
  • Azure AI Pipeline: https://github.com/owndev/Open-WebUI-Functions/blob/main/functions/azure_ai.py
  • N8N Pipeline: https://github.com/owndev/Open-WebUI-Functions/blob/main/functions/n8n.py
  • Google Gemini Pipeline: https://github.com/owndev/Open-WebUI-Functions/blob/main/functions/google_gemini.py
  • LiteLLM: https://github.com/BerriAI/litellm

Highlighted Details

• Azure AI: Seamless integration with Azure OpenAI, supporting Azure Search for RAG, multiple models, and customizable prefixes.
• N8N Automation: Connects Open WebUI to N8N workflows, featuring an AI Agent tool usage display (non-streaming only).
• Google Gemini: Advanced image processing with configurable quality/resizing, grounding via Google Search, and native tool calling.
• Security: Automatic, transparent encryption of API keys and sensitive data via WEBUI_SECRET_KEY.
• Monitoring: Time Token Tracker measures AI response times and token usage, with optional Azure Log Analytics integration.

Maintenance & Community

Maintained by owndev. No specific community channels or contributor details are listed in the provided README.

Licensing & Compatibility

Licensed under the Apache License 2.0, permitting broad commercial use and integration without significant copyleft restrictions.

Limitations & Caveats

The N8N pipeline's AI Agent tool usage display is restricted to non-streaming mode. Token counting accuracy is limited to OpenAI models via the tiktoken library.