# Rate limits & usage

Two limits apply to your account, both based on your subscription plan:

- A **monthly generation quota** that resets at the start of your billing period.
- A **per-minute request limit** on how often you can call the rendering endpoints.

| Plan    | Generations/month | Requests/minute |
| ------- | ------------------ | ---------------- |
| Free    | 50                  | 15                |
| Starter | 2,000               | 60                |
| Pro     | 5,000               | 150               |

Both limits apply to every endpoint that renders an image: `POST /api/generate`,
`POST /api/hooks/render`, and a fresh (uncached) `GET /api/render`.

> **Note: What doesn't count toward your limit**
>
> Cached responses never count toward either limit, on any endpoint. A render is only billed and rate limited the first time it's produced; repeat requests for the same template, layers, dimensions, and format are served from cache for free.

> **Note: GET /api/render is billed to the template owner**
>
> The public, unauthenticated `/api/render` endpoint is not exempt from rate limits. A fresh render on that endpoint counts against the monthly quota and per-minute limit of the **template's owner**, the account that created the template, not the caller. If you embed a template publicly and it gets popular, its owner's plan is what determines how much headroom you have.

## 429 responses

When either limit is exceeded, the endpoint returns HTTP 429 with a JSON body and a `Retry-After`
header:

- **Per-minute limit exceeded**: the body includes `limit` and `retryAfter` (seconds until the
	current one-minute window frees up).
- **Monthly quota exceeded**: the body includes `limit`, `used`, `remaining`, and `retryAfter`
	(seconds until your billing period resets).

In both cases, the `Retry-After` header mirrors the body's `retryAfter` value in seconds.