AnythingLLM

AnythingLLM is an open-source AI application from Mintplex Labs that runs fully on a user's own computer, letting people chat with documents, build no-code AI agent workflows, and use retrieval-augmented generation without sending data to the cloud. It ingests PDFs, Word files, CSVs, codebases, and other document types into local workspaces, then answers questions using either local models through Ollama or LM Studio, or connected APIs from providers like OpenAI, Azure, and AWS. Supported vector databases include LanceDB, PGVector, and Pinecone, and the tool is compatible with the Model Context Protocol (MCP) for connecting external tools to agents. Because everything can run offline and locally, AnythingLLM appeals to privacy-conscious individuals and to enterprises — the site names customers such as Merck, Oracle, Intel, and Samsung — that want document AI without exposing sensitive data to third-party servers. It ships as a free desktop app for Windows, macOS, and Linux under an MIT license, with an optional paid Cloud tier for team collaboration, plus a developer API for custom integrations. The tradeoff is setup: running everything locally with open models generally requires more configuration and hardware than a hosted, cloud-only competitor, and answer quality depends heavily on which model the user chooses to run.