AnythingLLM

Getting a private chat assistant that can read your own files usually means stitching together a model runner, a vector database and a front end.

AnythingLLM puts those pieces into one desktop app, so you can chat with your documents without writing code.

It is an open source AI assistant and agent tool that runs locally and privately, and it is free to use.

AnythingLLM works with various document types and supports custom models as well as enterprise models, including multi-modal ones.

You can extend it with agents and data connectors, and a built-in developer API lets you wire it into your own projects.

A growing set of plugins adds more.

The catch is hardware.

Running language models locally can strain your machine, and you have to download and install the desktop application first.

Self-hosting or cloud deployment may call for some technical knowledge.

If privacy matters more to you than convenience, that trade is reasonable, and the interface stays simple while you do it.