PrivateGPT is an open-source project for building private, retrieval-augmented AI applications that can answer questions using local documents. It is designed so files can be ingested, indexed, searched, and used as context for language models while keeping data inside a controlled environment. The project supports local components as well as configurable model and vector-store providers, allowing developers to choose how much of the stack remains fully offline.

PrivateGPT is commonly used as a reference architecture or starting point for document assistants, internal knowledge systems, and privacy-sensitive RAG deployments. Its APIs and modular design make it more developer-oriented than a simple desktop chatbot, while still demonstrating a complete ingestion-to-answer pipeline. It is especially useful for organizations and technical users who need AI-assisted document search but cannot freely send source material to external cloud services.