PrivateGPT
PrivateGPT is an open-source API layer that turns a locally running, OpenAI-compatible model server into a full private-AI application backend, modeled on the Claude API's shape.
It connects to any server implementing the OpenAI chat-completions and models endpoints, such as Ollama, llama.cpp or vLLM, but does not run models itself. On top of that connection it provides a standard messages API with streaming and token counting, file and artifact ingestion, retrieval with citations and agentic RAG, Claude-API-style built-in tools such as web search, web fetch and code execution, custom tools, MCP connectors, and structured access to databases and CSV files.
A bundled workbench UI at a local address is meant for testing, demos and quick local use, including sending messages, picking a model, uploading documents, testing retrieval citations and inspecting raw requests and responses, though the README describes the API itself as the intended product for other developers to build on. It is used in production inside Zylon, an on-premise AI platform for enterprises.
Python, Apache-2.0.