Vault AI (OP Vault) is a free, open-source web app that gives ChatGPT a long-term memory over your own documents. Upload your knowledge base — PDFs, text files, ebooks and more — and Vault embeds everything into a vector database, so you can chat with an assistant that answers from your files instead of only its training data.
Vault AI is an open-source project on GitHub with about 3,400 stars, built with Go and React on what the author calls the OP Stack: OpenAI plus the Pinecone vector database (with Qdrant support in the codebase). Documents are chunked and embedded when uploaded, and at question time the most relevant passages are retrieved and passed to the model as context, complete with a clean drag-and-drop web interface. The project became a reference implementation for the now-standard retrieval-augmented generation (RAG) pattern: many later chat-with-your-docs tools follow the same architecture. A hosted demo runs at vault.pash.city, and the whole stack can be self-hosted with your own API keys.
Vault AI is completely free and open source under the MIT license. Self-hosting is free apart from your own OpenAI API usage and any vector-database costs.
Vault AI is for developers and power users who want a ChatGPT-style interface over their own document collections, and for anyone learning how retrieval-augmented generation works in a real, readable codebase.