AnythingLLM1.16.1

AnythingLLM lets you chat with documents using a private AI assistant that runs entirely on your computer, keeping document data on the device. Each document becomes part of an agent knowledge base, supplying context for research, conversations, and workflow building. It runs without accounts, API keys, or token limits, and the desktop edition works on macOS, Windows, and Linux; the Mac release targets Apple Silicon (arm64) and is currently version 1.16.2.

Its Meeting Assistant records, transcribes, and summarizes meetings locally, producing full transcripts, action items, and key decisions without a bot joining the call or sending audio for cloud processing. It works with any meeting platform and indexes meeting content into the assistant’s context. A pointer-based assistant can refine spoken words in real time while using what is on screen to provide more precise answers.

Users can select a preferred model or let AnythingLLM recommend one based on their hardware specifications, and the desktop software is packaged as one file with no dependencies. A custom tool builder and developer API support tailored workflows, while You.com real-time web search is enabled from the first chat without an API key or setup. AnythingLLM Mobile runs small models on the device and includes web search, document chat, file generation, scheduled jobs, and memory; pairing it with the desktop edition provides access to larger models and full workspaces.

AnythingLLM is freeware, and the project is MIT-licensed open source for organizational use, with self-hosted multi-user deployment available on private servers alongside the desktop app.