PrivateGPT, from Zylon, is an open-source project that provides a full API layer for building private AI applications on local models. It includes retrieval-augmented generation, skills, tools, MCP support and text-to-SQL, among other features.
It works with any inference server that is compatible with the OpenAI API, so you can choose how your models are run. The project is free on GitHub and has a very large following.
