Haervwe/open-webui-tools

What it solves

This repository provides a modular collection of over 20 specialized tools, function pipes, and filters designed to extend the capabilities of Open WebUI. It transforms a standard AI chat interface into a versatile workstation capable of academic research, multi-modal content generation, and autonomous agentic behavior.

How it works

Users can install these components via the Open WebUI Hub or by manually adding the Python scripts to their Open WebUI instance. The collection is divided into three main categories:

• Tools: Extensions that allow the LLM to perform specific actions, such as searching arXiv for academic papers, querying YouTube, or generating images via Hugging Face and Cloudflare.
• Function Pipes: Advanced logic layers that enable complex behaviors, including autonomous planning (Planner Agent), Monte Carlo Tree Search for research, and multi-model conversations.
• Filters: Middleware that modifies inputs or outputs, such as a prompt enhancer, a semantic router for model selection, or a tool that cleans "thinking tags" from responses.

Who it’s for

This toolkit is for users of Open WebUI who want to add professional-grade functionality to their local AI setup, specifically those interested in academic research, AI-driven content creation (image, video, and music), and autonomous agent orchestration.

Highlights

• Academic Research: Integrated arXiv search and MCTS-based research pipes.
• Multi-modal Generation: Tools for text-to-image, text-to-video, and music generation via ComfyUI, Google Veo, and Atlas Cloud.
• Autonomous Agents: A Planner Agent with multi-agent delegation and real-time execution tracking.
• Web Integration: Search capabilities via Perplexica, SerpBase, and SearxNG, plus X (Twitter) data retrieval.
• System Utilities: Tools for checking provider health (OutageDeck) and managing local music servers (Mopidy).