Make the AI your own.
Choose how your coding model runs, and know exactly what it costs.
Coding model
A real inference connection, not a simulated chat.
The model is managed by the platform; its name, license, and open-source status are not disclosed. Free access is not unlimited: 100 requests per hour per app, up to 8,192 output tokens per call, and a 60-second request timeout. Limits are shared across users. No live web search on this plan.
Connected accounts & services
Connect your own accounts using managed authorization. Actions may require a paid plan.
Sign in to save and reopen your private projects across devices.
Private project storage
Files, conversations, checkpoints, and settings use account-scoped key-value storage.
Accounts are handled by the built-in authentication service. Project keys are isolated server-side by account, not by a client-side role or ID. Public uploaded media and published assets are not private. Never upload credentials, payment details, or private documents.
Sign in to save and reopen your private projects across devices.