Pricing & billing
The model
Section titled “The model”Developer access is part of Pro ($25/month). Usage — AI generation and hosting — is billed pay-as-you-go from a USD wallet at transparent, published USD rates. No credits, no seats, no tiers.
- Wallet: prepaid USD balance (
GET /v1/billing/wallet, top up viaPOST /v1/billing/topupor the billing page). - Ledger: every debit is a line item —
GET /v1/billing/ledgershows the amount, the surface (generation/hosting), the app, and your balance after. - Live prices:
GET /v1/billing/pricingalways returns the current price list; the numbers below are a snapshot of that endpoint.
AI generation
Section titled “AI generation”AI generation is metered by tokens, at these published USD rates:
| Meter | Price |
|---|---|
| Input tokens | $1.80 / 1M |
| Output tokens | $5.72 / 1M |
| Cached input | $0.34 / 1M |
In practice: creating an app costs ~$1.50–2.50, an edit ~$0.50–1.50 (edits get cheaper as the cache warms). Each operation reserves an estimate and settles to actual on completion; overage past the estimate cap is absorbed, and failed generations are not charged.
Hosting & calls
Section titled “Hosting & calls”| Meter | Price |
|---|---|
| API / MCP / A2A / gateway calls | $1.00 / 1M calls |
| CPU-ms past the included floor | $0.10 / 1M |
| Requests, database rows, file-storage operations, emails | metered per app — see GET /v1/apps/{id}/hosting-usage |
Spend controls
Section titled “Spend controls”- Hard monthly cap:
PUT /v1/billing/limits{"monthly_spend_limit_usd": 60}— operations that would exceed it are refused (nullclears it). - Auto-reload:
POST /v1/billing/auto-reloadto top the wallet up automatically. - Usage rollup:
GET /v1/billing/usage— spend by surface for the current period.
When the wallet can’t cover an operation you get 402 Payment Required with error code insufficient_balance — top up or raise your cap, nothing is silently queued.