PrivateGPT

PrivateGPT is an open-source API layer, built by Zylon and following the Claude API model, that turns local LLMs into production-ready private AI applications. It provides higher-level building blocks such as a messages API with streaming, document and artifact ingestion, RAG retrieval with citations, embeddings and orchestration, and custom tools/MCP, so all data stays within your own environment. It ships a built-in workbench UI and works with any OpenAI-compatible local inference server like Ollama, llama.cpp, vLLM, and LocalAI.