AnythingLLM

AnythingLLM is an open-source application for turning your own documents into a chat interface powered by retrieval-augmented generation. It runs as a desktop app on Mac, Windows, and Linux, or as a self-hosted Docker server, and it connects to whatever language model you prefer: local models through Ollama or LM Studio, or hosted ones from OpenAI, Anthropic, Azure, Gemini, and others. Everything, including the vector database and document processing, can stay on your own machine, which is why it's popular with people and teams who don't want their files sent to a third party.

The core organizing unit is the workspace: each one holds its own set of uploaded documents, embeddings, and chat history, so a legal team's contracts and a developer's codebase can sit in the same install without their context mixing. It supports a wide range of file types (PDFs, Word docs, code, CSVs, web pages) and can also pull in data from tools like Confluence, GitHub, or YouTube transcripts. Beyond basic Q&A, it includes agent features, letting a workspace call web search, scrape a page, or run other tools mid-conversation, and it exposes a developer API for embedding the same RAG pipeline into other applications.

What sets it apart from hosted chat-with-your-docs services is the combination of full local control and model flexibility: you're not locked into one LLM provider or forced to upload sensitive files to someone else's cloud. It's aimed at developers, small businesses, and privacy-conscious individuals who want a ChatGPT-like experience over private data without giving up ownership of that data or paying per-seat SaaS fees.