> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bridgly.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Rate limits

> Default limits, how they can vary by tier or custom agreement, and what happens when you exceed them.

Bridgly limits how quickly you can send requests to keep the API reliable for
everyone.

## Default limits

**Your actual rate limits may vary based on your tier and any custom agreement.**
The defaults for authenticated API requests are:

| Window     | Requests |
| ---------- | -------- |
| Per second | 5        |
| Per minute | 100      |

Both limits apply. Staying below 5 requests per second can still exceed the
100-request minute limit.

Limits are applied **per account, per endpoint** and shared across all API keys
belonging to that account. Creating another key does not increase your limit for
the same endpoint. Requests made through the SDK or MCP server count toward the
same limits as direct API calls.

## When a limit is exceeded

The API rejects the request with **`429 Too Many Requests`**. The scrape does not
run, and the rejected request does not consume credits.

The response includes a `Retry-After` header with the number of seconds to wait
before trying again. For example:

```http theme={"dark"}
HTTP/1.1 429 Too Many Requests
Retry-After: 10
Content-Type: application/json

{
  "status": "error",
  "message": "Too many requests. Please slow down.",
  "retryAfterSeconds": 10
}
```

Wait at least the indicated delay, then retry at a slower pace. Spread requests
out instead of sending large bursts, and reduce parallel requests if you keep
receiving `429` responses.

## Response headers

The API also reports the shortest configured rate-limit window through:

| Header                  | Meaning                                  |
| ----------------------- | ---------------------------------------- |
| `X-RateLimit-Limit`     | Maximum requests allowed in that window. |
| `X-RateLimit-Remaining` | Requests remaining in that window.       |
| `Retry-After`           | Seconds to wait after a `429` response.  |

The remaining count covers that window only; a longer window may still be
exhausted. See [Responses & errors](/responses) for the other API status codes.