Everything you create through the API or a connected assistant uses the same credits as the Samsa app — there's no separate API balance.
API and MCP activity draws from your organization's monthly credit pool — the same pool the app uses — and it respects any team budgets you've set up. A generation costs the same number of credits whether it's made in the app, through the API, or by a connected assistant. Read-only actions — listing models, checking a job's status, or reading your balance — are free.
If your organization runs out of credits, requests that cost credits will fail until the pool is topped up or resets. For how the shared pool and budgets work, see Shared Credits and Teams & Credit Budgets.
You can see exactly what your keys are doing:
API Keys tab — in Organization Settings > API Keys, each key shows its requests, credits used (this period / last 30 days / all time), when it was last used, and a breakdown of credits by action type.
Analytics — open the avatar menu in the bottom-left corner and choose Analytics; the API section shows API usage across your organization.
Different actions cost different amounts of credits — a video costs more than a single image, and higher resolutions cost more. The API uses the same pricing as the app. For the current per-action pricing, see the Pricing guide in the developer docs (opens in a new tab).
To keep the service fast and reliable for everyone, the API applies rate limits — a cap on how many requests you can make in a short window. If you go over, the API asks you to wait a moment and try again. The current limits and how to handle them are in the Rate limits guide (opens in a new tab).
If a request can't go through — for example a missing scope, not enough credits, or too many requests — the API returns a clear error explaining why. The Errors guide lists what each one means and how to respond (opens in a new tab).
