Quick start

From signing up to holding your first delivery usually takes about five minutes.

1. Create an account

Sign up with an email address, or with GitHub / Google. Verify your email to receive $1 of one-time model usage credit — enough to run a few small projects through from end to end.

2. Connect a computer

Tasks run on a computer of your own (macOS / Linux / Windows WSL). In the app, click Add computerand paste the generated one-line install command into that machine’s terminal: the client registers itself and comes online, and the page lights up as it becomes available. Credit and top-ups are spent on model usage only — you bring the compute.

3. Send your first requirement

In Projects, create a project and describe the result you want in one sentence — say, “a page for collecting weekly team reports, with CSV export”. The AI team comes back with a plan and a design first, then starts building real source code.

4. Follow along and steer

The whole run is visible in the workspace: conversation, tasks and artifacts accumulate as they happen. You can step in at any time to change direction, and the team carries its context forward instead of starting over.

5. Collect the delivery

Once the review gate passes, the project’s Delivery page produces a signed delivery card: complete source code, build artifacts and a deployment path (hosted release on the platform, or your own deployment target). What you get is a shippable product, not a chat log.

How billing works

You pay for model usage, 1 credit = $0.01, itemised under Settings → Usage in real time. The free tier has no monthly cloud allowance — when the starter credit runs out, top up or subscribe to Pro to continue. See pricing.