AnythingLLM is an all-in-one desktop application for Mac, Windows, and Linux that enables you to download and run any LLM locally and privately chat with your documents. By default, the LLM, embedding, and even vector database are all stored locally and can even run 100% offline.

If you are looking to leverage more powerful models, embedders, or vector databases you can connect to any of the 20+ integrated providers to supercharge your AnythingLLM instance.

Shipped with Ollama built-in, you do not need to download anything else to get a full RAG chatbot with zero setup and full privacy. Upload PDFs, mp3s, text files - Anything!

AnythingLLM also comes with a full developer API you can leverage for custom integrations and even several data source connectors like GitHub and YouTube.

AnythingLLM is an MIT Licensed project and is fully open-source for you to modify.

If you are looking for a multi-user experience AnythingLLM also has a self-hosted Docker image that supports concurrent users. AnythingLLM is the last RAG-based chatbot solution you will ever need.