Quickstart: build an app

1. Open Dashboard → Build, connect your wallet and choose New app. Describe what users can do, what belongs on chain and what the page should show.
2. Attach a long specification with the paperclip. Long pasted briefs become text files with a visible attachment chip. The planner receives attached text before it decides whether the app needs contracts or a wallet. Add a short instruction such as 'Use the attached brief.'
3. Choose Platform model or a saved model connection. Read the cost on the Build button and send. Enter sends; Shift+Enter inserts a newline. While Enhance is running, wait for it to finish before submitting.
4. If both execution slots are busy, the reply shows that your build is queued and its position. Cancel removes a queued run. Edit restores its prompt, files and model selection to the composer. Once it starts, Build details streams the recorded steps.
5. Follow Idea → Contract → Web page → Published. Read the toolchain verdict above the workspace. Open the live URL once a publication is recorded, then test the app's own interactions on the stated network.

A raffle where anyone buys tickets with ETH, the owner draws a winner after a block deadline, and the winner claims the pot.

• Open Build

A contract build runs compilation and tests in a sandbox, then deployment and hosting when those stages succeed. Good builds take time because each reported result has to be checked. A front-end-only app follows the stages it actually needs. The compiler verdict stays visible even when the model's reply is shorter.

Closing the browser leaves the server-owned build running. Reopen the project to reconnect to its stored progress. A failed charged build receives its Fletch credit refund; provider charges for own-key usage follow the provider's billing. Read the current execution limits before starting a large specification.

• Credits, time limits and current availability