Magibu LLM Tools
Tools and auxiliary systems the model uses at inference time - overcoming model limits by calling the right tool at the right time.
Vision
Build a modular, independently testable tool ecosystem so a language model can reliably do what it cannot do alone (exact math, fresh information, OCR, domain RAG). Each tool has a clear interface and runs deterministically or in isolation.
What it does
Provides ready tools like string operations, math, a Python sandbox, and web search. Translate, OCR, Medical RAG, and expert model routing are planned.
Architecture
Question → LLM decision → Tool Router → appropriate tool → response. Each tool: clear schema + independence + unit tests.
Contribution areas
- Build new tool modules (+ tests)
- Translate or OCR expert model integration
- Medical RAG: cited article indexing pipeline
- Tool router evaluation infrastructure
Tech stack: Python · LLM Inference · Sandbox · MCP
I want to join this project
Verify your Google account, fill out the form, then pick a task from the GitHub issue list to get started.