Rate limits

The provider can answer with HTTP 429 when account or channel limits kick in. The dashboard key API does not set RPM, TPM, concurrency or a budget for a key. Below: what the key settings lack, and how to handle 429 in a client.

Per-key quota settings

The key API does not publish numeric default quotas and does not offer these settings:

Per-key configuration

A key can have a name, a Premium or Cheap channel, a Premium fallback for Cheap and an expiry date, and its usage appears in the statistics. RPM/TPM, concurrency, budgets, model lists, IP filters and IDE restrictions cannot be set through the dashboard key API.

Supported settings:

• Name and Premium or Cheap channel
• Optional Premium fallback for a Cheap key
• Optional future expiry

Rate limit headers

If the provider sends these headers, rely on them when retrying. The key API does not guarantee that they are present:

Handling 429 errors

When the limit is reached, the API answers with HTTP 429. Retry with an exponentially growing delay:

import time
import requests

def request_with_backoff(url, headers, data, max_retries=5):
    for attempt in range(max_retries):
        response = requests.post(url, headers=headers, json=data)
        if response.status_code == 429:
            wait = 2 ** attempt  # 1, 2, 4, 8, 16 seconds
            print(f"Rate limited. Retrying in {wait}s...")
            time.sleep(wait)
            continue
        return response
    raise Exception("Max retries exceeded")

async function requestWithBackoff(
  url: string, options: RequestInit, maxRetries = 5
) {
  for (let attempt = 0; attempt < maxRetries; attempt++) {
    const res = await fetch(url, options);
    if (res.status === 429) {
      const wait = 2 ** attempt * 1000;
      console.log(`Rate limited. Retrying in ${wait}ms...`);
      await new Promise(r => setTimeout(r, wait));
      continue;
    }
    return res;
  }
  throw new Error("Max retries exceeded");
}

Best practices

• If the headers are there, check X-RateLimit-Remaining and Retry-After before retrying.
• Named keys help you account for workloads separately, but they do not create separate quotas.
• Retry with exponential backoff plus jitter.
• Cap automated workloads on the client side and watch customer cost. Key budgets cannot be set through this API.