Skip to content
aarness.

Documentation

Learn the harness

Run Haarness on your machine, script it, extend it, connect any model and take it to the cloud.

Start here

  1. 1What Haarness isHaarness is a coding agent that works in your project from the terminal, a desktop app or a script, with the model you choose. Start here to install it, connect a model and run your first chat.
  2. 2Install HaarnessInstall the haarness command and the Desktop app on macOS, Linux or Windows with one command, check that it works, update it, pick a preview or a specific version, and uninstall it.
  3. 3Connect a providerConnect a model to Haarness by signing in with ChatGPT, GitHub Copilot, xAI or Haarness Cloud, pasting an API key, or using Ollama or LM Studio, from the terminal UI, Desktop or the CLI.
  4. 4Your first chatOpen Haarness in a project, ask a question, let it make a change, answer a permission prompt, review and undo the change, stop a run, and resume the chat later, step by step in the terminal UI.

Every section

Start here5 pages

What Haarness is, installing it, connecting a provider and your first chat

Terminal UI8 pages

Chats in the terminal, review, goals, slash commands and keyboard shortcuts

Desktop6 pages

Installing the app, chats, reviewing changes, previews and settings

Haarness CLI14 pages

Scripted runs, resuming chats, worktrees, settings, extensions and schedules

Server & SDKs5 pages

Starting a server, connecting safely, the API, ACP and the SDKs

Models & providers5 pages

Choosing a model, sign-ins and API keys, local models and the Router

Permissions & security4 pages

Permissions and approvals, budgets, trust boundaries and your data

Haarness Cloud19 pages

Cloud workspaces, agents, schedules, remote control, credits and self-hosting

Still looking?

Ask us, see what changed lately, or find a plugin that already does it.