Introduction
PrivateGPT is a production-ready, privacy-first project that provides tools and an API for private document question-answering and Retrieval-Augmented Generation (RAG). It is designed to run in offline or private-cloud environments so that data never leaves the execution environment.
Key features
- Privacy-first architecture: designed for offline and air-gapped deployments.
- End-to-end RAG pipeline: document ingestion, chunking, embedding generation and vector retrieval.
- Multiple backends: compatible with LlamaIndex, Qdrant, Ollama and other runtimes and vector stores.
- Deployment-ready: Docker Compose, cloud and on-premise deployment examples and scripts.
Use cases
- Private knowledge bases and document Q&A in enterprises.
- Regulated industries (healthcare, legal) requiring strict data isolation.
- Offline or network-restricted research and production deployments.
Technical details
- FastAPI-based OpenAI-compatible API with LlamaIndex powering the RAG components.
- Primarily Python-based, documentation and MDX/Gradio UI included for demos and examples.
- Apache-2.0 license; active community and release cadence.