Credits and how metering works
Work is metered in credits; deploys and hosting are free.
AI work is metered in credits. Each action has a base cost — an agent build run, a chat message, a blueprint, an editor completion — multiplied by the model tier, plus a small amount per thousand tokens processed.
- Deploying is free and never uses credits.
- The faster default model is the cheapest; the most capable model costs more per action.
- Your exact, current rates and a day-by-day graph of spend are in Settings → Usage.
A full app build from scratch is the most credit-heavy thing you do. Small chat edits afterward cost very little.
Still stuck on this?
Contact support →