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

> ## Agent Instructions
> Krun API base URL: https://api.krun.ai. Authenticate server-side with `Authorization: Bearer $KRUN_API_KEY`; never put a Krun API key in browser code.
> Prefer the Python SDK in examples: `pip install krun-ai`, `from krun import Krun`. The TypeScript SDK `@krun-ai/sdk` is not published on npm yet: do not tell users to `npm install` it.
> The only question type is `choice`. `confidence` is the top-1 minus top-2 probability margin, not the probability that the answer is correct. `choice` is null when `abstain` is true.
> Usage reports `input_tokens` only. There are no output tokens.

# Rate limits and quotas

> How per-key rate limits and per-project monthly quotas work in the Krun API, and how to handle 429 errors.

Krun applies two kinds of limits:

|                | Scope       | Window               | Error                |
| -------------- | ----------- | -------------------- | -------------------- |
| **Rate limit** | Per API key | Requests per minute  | `429 RATE_LIMITED`   |
| **Quota**      | Per project | Calendar month (UTC) | `429 QUOTA_EXCEEDED` |

Limits are set per project and per API key during the closed beta, so they can differ between projects. Contact [support](/resources/support) if you need higher limits.

## Rate limits

Each API key can make a limited number of requests per minute, counted in fixed one-minute windows. Every response includes the current state:

| Header                  | Description                               |
| ----------------------- | ----------------------------------------- |
| `X-RateLimit-Limit`     | Requests allowed per window for this key. |
| `X-RateLimit-Remaining` | Requests left in the current window.      |
| `X-RateLimit-Reset`     | Seconds until the window resets.          |

When you exceed the limit, the API returns `429` with `RATE_LIMITED` and a `Retry-After` header with the number of seconds to wait:

```json theme={null}
{
  "error": {
    "code": "RATE_LIMITED",
    "message": "...",
    "request_id": "req_..."
  }
}
```

Wait for `Retry-After` before sending more requests. To stay under the limit, ask several questions about the same text in [one request](/guides/multiple-questions) instead of one request per question.

## Quotas

A project can have monthly limits on:

* **decisions**: each question in a successful request counts as one decision,
* **input tokens**: the sum of `usage.input_tokens` of successful requests.

Only successful decisions count. When a quota is exhausted, the API returns `429` with `QUOTA_EXCEEDED` until the next calendar month (UTC). Retrying doesn't help.

<Note>
  Quotas are checked before each request, so concurrent requests can go slightly over a quota.
</Note>

## Handling 429 in Python

```python theme={null}
import time

import krun

try:
    result = client.decide(context="...", questions={...})
except krun.RateLimitError as e:
    time.sleep(e.retry_after or 1)
    result = client.decide(context="...", questions={...})
except krun.QuotaExceededError:
    ...  # stop sending requests and alert your team
```