Getting started

Install tomo and stand up a personal AI agent on your own machine: a language model behind your chat apps, acting through gated tools, pointed at any provider you choose.

Three short pages: how tomo thinks about an agent (a local daemon behind your chat apps, acting through tools that pass a policy gate first, pointed at whatever model you like), how to install the single binary, and a guided first run that ends with a working agent you can talk to in the terminal and in a browser.

tomo (友, "companion") runs on your hardware. Your provider key, your history, and your memory stay on the box, and the only thing that leaves is the model call you would have made anyway.

Start with the introduction for the shape of it, then install and follow the quick start.

Introduction What tomo is, why a local-first agent keeps your keys and history on your machine, how the gate sits below the model, and when you would reach for it. Installation Install tomo from Go, a release archive, a Linux package, Homebrew or Scoop, or the container image, and what it needs to run. Quick start From an empty terminal to a working agent: set a key, onboard, talk to tomo in the terminal, then open the local web chat.