Rate limits and credit / spending caps for the Pioneer API

Per-endpoint request-rate limits, edge WAF quotas, monthly credit and overage spending caps, 429 handling, and how to request higher limits on the Pioneer API.

Request-rate quotas and credit / spending caps for the Pioneer API, how to handle 429 errors, and how to request higher limits

The Pioneer API enforces two independent things that can stop a request: request-rate limits that cap how many API calls you can make per minute or hour, and credit-based usage limits that cap how much you can spend. Exceeding a request-rate limit returns 429 Too Many Requests. Running out of credits or hitting your plan’s overage ceiling returns 402 Payment Required or 403 Forbidden instead — see Credit limits and overage spending cap below.

Request-rate limits

Two independent layers protect the API:

1. Edge rate limit — always applied to every request at the load balancer, before it reaches the API, regardless of endpoint or authentication. Aggregated by the IP address the edge observes, which is not always your application’s true client IP (for example, requests proxied through a shared egress hop are aggregated together). Limit: 100,000 requests / 60 seconds.
2. Per-endpoint limit — most endpoints below enforce their own limit scoped to your billing team (falling back to API key, then user, then client IP for unauthenticated requests). This is the limit that governs a normal, authenticated caller. It replaces the generic per-IP default for that endpoint rather than stacking on top of it — the per-IP default only governs endpoints with no listed override.

Endpoint	Scope	Limit
All other endpoints (no endpoint-specific limit set)	Per client IP	20,000 / min · 1,000,000 / hour
POST /inference	Per user	5,000 / min
POST /v1/chat/completions, /v1/completions, /v1/responses, /v1/messages	Per user	5,000 / min
POST /gliner-2/*	Per user	15,000 / min
POST /generate/*	Per user	120 / min
POST /felix/training-jobs	Per user	20 / min

For a single API key or team, the per-endpoint limit above is what actually binds. The 100,000 requests / 60 second edge limit is a separate, always-on ceiling shared by all traffic through the same load balancer — it only comes into play when many different callers share the same observed IP and collectively exceed it.

Credit limits and overage spending cap

Inference is billed against a credit balance rather than a request-rate window (1 credit = $0.01). Each plan includes a credit allowance — the Free plan grants a one-time allowance that does not renew, while paid plans renew their included credits every billing month. Once a paid plan’s included credits are used, additional usage draws from overage billing (if enabled) up to that plan’s maximum overage spend per month; on the Free plan, running out simply stops inference until you add credits or upgrade.

Exceeding a credit limit does not return 429 Too Many Requests. Instead it returns:

• 402 Payment Required when your included credits are exhausted and there’s no spendable balance to draw from (code: "out_of_credits").
• 403 Forbidden when your plan’s maximum monthly overage spend has been reached (code: "credit_ceiling_reached").

Both responses share the same JSON shape:

{
  "code": "out_of_credits",
  "message": "You've used your included credits. Add credits or enable auto top-up.",
  "resolution_url": "https://agent.pioneer.ai/credits"
}

You can check your current usage, remaining allowance, and overage settings anytime in the billing section of the dashboard.

Credit limits, overage ceilings, and plan terms are subject to availability and may be adjusted over time.

Need a higher limit? Reach out to [e-mail] or your account contact and we can raise the ceiling on a custom plan.

Handling 429 responses

When you exceed a limit, the API returns 429 Too Many Requests and includes a Retry-After header that tells you how many seconds to wait before retrying.

cURL

HTTP/2 429
retry-after: 3
content-type: application/json

{
  "detail": "Rate limit exceeded: ..."
}

The following pattern handles 429 responses with a simple sleep-and-retry loop:

Python

import time
import requests

def call_with_retry(url, headers, payload, max_retries=5):
    for attempt in range(max_retries):
        response = requests.post(url, headers=headers, json=payload)

        if response.status_code == 429:
            retry_after = int(response.headers.get("Retry-After", 1))
            print(f"Rate limited. Retrying in {retry_after}s...")
            time.sleep(retry_after)
            continue

        response.raise_for_status()
        return response.json()

    raise RuntimeError("Max retries exceeded.")

Credit and overage denials (402/403, see Credit limits and overage spending cap) won’t resolve by waiting — the retry loop above only applies to 429 responses. A 402/403 needs a billing action (add credits, enable auto top-up, or upgrade your plan) before the next request can succeed.

Requesting higher limits

If the default or Pro-tier limits don’t fit your workload, contact the Pioneer team to discuss a custom plan.

Request higher limits