Questions, answered plainly.

A desktop app for running teams of AI coding agents. It brings its own local coordination layer, so agents can hold roles, claim tasks, message each other and review each other's diffs. You supervise all of it from one kanban board instead of several terminal windows.

Install the app, open a project, and create a team with a few roles. Give the team an objective. The lead breaks it into tasks and assigns them, builders work in parallel, and the reviewer checks the diffs. You watch the board and step in when a card reaches review.

Yes. Solo mode gives you a single agent that writes and works its own task list on the same board. When the work outgrows it, add roles and that agent becomes the team lead without losing any history.

Use them — AgentTeams runs on top of them. What it adds is the layer the CLIs leave out: parallel agents with roles, a board that shows state, messages between agents, a diff review UI, and budgets. If you only ever run one agent, the CLI is enough.

macOS on both Apple Silicon and Intel, Windows on x64 and ARM64, and Linux via AppImage. The download page detects your platform and puts the right build first.

Start with three: a lead, a builder and a reviewer. Add builders only when the board shows cards waiting in assigned rather than in review. Most people over-provision builders and end up bottlenecked on review instead.

Give each teammate its own workspace and they never touch the same working tree. For work that genuinely overlaps, link the tasks so one blocks the other and the board enforces the order.

Claude Code, Codex and OpenCode run as first-class runtimes. Cursor, GitHub Copilot, SuperGrok, Z.AI, MiniMax, Kimi and Kiro connect through supported provider routes, and OpenRouter opens up 200+ models across 75+ providers. Local models work through Ollama.

Yes, and it is usually the right call. Put a strong reasoning model on the lead and reviewer roles, and a cheaper or faster model on the builders. Cost drops sharply and quality holds because review is where mistakes get caught.

Coordination state, task history and app data stay on your machine. Code only leaves your machine if you point an agent at a hosted model — and you choose that per team, so you can keep sensitive repos on a local runtime.

Yes. The app is free and open source, and it ships with a free model that needs no authentication — no account, no API key, no card. Paid providers are entirely optional and only apply if you connect one.

Set a monthly ceiling in tokens or currency. You get an alert at 80% and again at 100%, and scheduled runs can hard-stop at the cap rather than continuing. Spend is broken down per team, agent, task and model so you can see what is expensive.