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

# Rate limits

> Organization-level rate limits, response headers, and 429 handling on the Seamless API.

Every endpoint is rate limited per organization. The limit is operator-configured and can differ by endpoint and by organization, so do not hardcode a number: read **`X-RateLimit-Limit`** on each response, which always reflects your actual limit for that endpoint in the current window. The standard allowance is on the order of 60 requests per minute per endpoint. When you exceed the limit, the API returns **429 Too Many Requests** with a JSON error body.

All API keys and users in your organization share the same quota. The limit applies separately to each endpoint — for example, search and poll each have their own window.

> [!TIP]
>
> **Need higher rate limits?** If you need higher custom rate limits, contact us for pricing details. [Contact sales →](https://seamless.ai/pricing-inquiry)

## Response headers

| Header | Example | Description |
| - | - | - |
| `X-RateLimit-Limit` | `60` | Max requests allowed in the current window for this endpoint |
| `X-RateLimit-Remaining` | `42` | Requests remaining in the window |
| `X-RateLimit-Reset` | `1745587198` | Epoch seconds when the window resets |
| `X-PublicAPI-Credits` | `1000` | Credits remaining for your org. Sent on every authenticated response. On a contact research `202`, it's your projected balance after every accepted contact is charged, using the most each contact can cost (see [Contact research costs](https://docs.seamless.ai/core-concepts/credits.md#contact-research-costs)). |

## Handle 429 errors

**1. Read the reset time**

Parse `X-RateLimit-Reset` and wait until that timestamp before retrying
the same endpoint.

**2. Backoff between polls**

When polling research results, use a 2–5 second interval instead of
tight loops.

**3. Avoid duplicate research**

A `duplicate` poll status means the record was already researched — do
not resubmit research for the same `searchResultId`.

```python theme={null}
import time
import requests

response = requests.get(
    "https://api.seamless.ai/api/client/v1/contacts/research/poll",
    headers={"Token": API_KEY},
    params={"requestIds": request_id},
)

if response.status_code == 429:
    reset = int(response.headers.get("X-RateLimit-Reset", 0))
    wait_seconds = max(reset - time.time(), 1)
    time.sleep(wait_seconds)
```

## Related

* [Credits](https://docs.seamless.ai/core-concepts/credits.md) — what uses credits and how to check your balance
* [Error codes](https://docs.seamless.ai/api-catalog/error-codes.md) — HTTP 400, 401, 403, 422, 429, and more
* [Troubleshooting](https://docs.seamless.ai/troubleshooting/authentication-and-request-failures.md) — failed research and auth errors


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