GH Repository · zylon-ai
private-gpt
Complete API layer for private AI applications on local models: RAG, skills, tools, MCP, text-to-sql, and more. Works with any OpenAI-compatible inference server.
- stars
- 57,517
- 30-day movement
- +52/day
- Related entries
- 11
- Connections
- 2
makepython/uvpythondockerotheron-premisePythonai-toolsai
private-gpt is a Python-based API layer for building private AI applications on locally hosted models. It exposes RAG, skills, tools, MCP, and text-to-sql capabilities, and works with any OpenAI-compatible inference server.
You reach for it when you need a ready-made API layer for private, on-premise AI features without wiring RAG and tooling yourself.
Use it to
- Serve RAG over local models via an API
- Add text-to-sql to private data
- Expose skills and tools through MCP
- Run AI features fully on-premise
- Point it at any OpenAI-compatible server
For Developers building private, self-hosted AI applications
- Role
- other
- Language
- Python
- Licence
- Apache-2.0
- Forks
- 7,616
- Open issues
- 1
- Last push
- 2026-09-16
- Latest release
- v0.0.1 · 2023-10-20
topicsraglocal-modelson-premiseapimcptext-to-sql