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

# Rate limits

> Understand MCP request quotas and retry behavior.

Zero applies a shared fixed-window limit to authenticated MCP transport requests. The default is **120 requests per 60 seconds per OAuth token or API key**.

Every authenticated response includes:

| Header | Meaning |
| - | - |
| `X-RateLimit-Limit` | Maximum requests in the current window. |
| `X-RateLimit-Remaining` | Requests remaining before the limit is enforced. |
| `X-RateLimit-Reset` | Unix timestamp when the window resets. |

## Exceeded limits

Requests over the limit receive `429 Too Many Requests` before the MCP operation runs:

```http theme={null}
HTTP/1.1 429 Too Many Requests
Retry-After: 18
X-RateLimit-Limit: 120
X-RateLimit-Remaining: 0
X-RateLimit-Reset: 1787191260
Content-Type: application/json

{"error":"MCP request rate limit exceeded"}
```

Wait for `Retry-After` before reconnecting or retrying. Add jitter when multiple agents share one credential.

## Limit precedence

Organization administrators can configure a different MCP quota. API-key-specific limits take precedence over organization limits, which take precedence over the global default. OAuth connections use the selected organization's quota.

REST API requests made by `execute` are also subject to the relevant [API rate limits](https://docs.zdlake.com/api/rate-limits.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.