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

# Rate limits

> How Sayless rate-limits public API traffic and which headers to read.

Every response includes standard rate-limit headers. Integrations should honor them before retrying.

## Headers

| Header | Meaning |
| - | - |
| `RateLimit-Limit` | Max requests in the current window |
| `RateLimit-Remaining` | Remaining requests in the window |
| `RateLimit-Reset` | When the window resets (ISO 8601) |
| `RateLimit-Violation-Count` | Consecutive violations (when present) |

On HTTP **429**:

| Header | Meaning |
| - | - |
| `Retry-After` | Seconds to wait before retrying |

```http theme={"system"}
HTTP/1.1 429 Too Many Requests
Retry-After: 120
RateLimit-Limit: 1000
RateLimit-Remaining: 0
RateLimit-Reset: 2026-09-30T14:32:00.000Z
```

API/OAuth error bodies use the nested [error envelope](/developers/errors); the payload may also include `retryAfter` / violation metadata depending on the path.

## Defaults

* **Default:** about **1000 requests per minute** per authenticated principal (user+IP) or IP for unauthenticated probes.
* Individual routes may declare stricter decorators (for example auth email send, or high-cost writes).
* Workspace plan-based hourly caps exist for some product features; curated Broad traffic is primarily gated by the per-route / default limiters above.

Backoff with jitter after 429s. Do not spin on `Retry-After: 0`.