Rate Limiting

How Rate Limiting Works

Router9 governs usage with two independent mechanisms per Token Plan:

• 4-hour window — A short burst guard that limits requests within any 4-hour period. Enforced on every request; exceeding it returns 429.
• Monthly credit quota — Governs your total usage over the billing month, weighted by per-model cost, and carries over 80% of what you don't use (up to 2× quota). See Credits & Usage.

The 4-hour window is the only enforced request-count limit; overall volume is otherwise metered by credits rather than a weekly request cap.

Rate Limits by Plan

Free plan model access: FREE accounts can only call models flagged as free-tier. Calls to premium models (e.g. Claude Sonnet, GPT-5) return 403 model_requires_upgrade — upgrade to Pro or Max for full model access.

Rate Limit Headers

Every API response includes the current 4-hour window state:

X-RateLimit-Limit-4h: 1000
X-RateLimit-Remaining-4h: 987
X-RateLimit-Reset-4h: 2026-06-04T16:00:00Z

Token endpoints also return your monthly credit state (X-Credits-Limit, X-Credits-Remaining, X-Credits-Reset) — see Credits & Usage.

429 Response

When a rate limit is exceeded:

HTTP/1.1 429 Too Many Requests
Retry-After: 1800

{
  "error": "4-hour rate limit exceeded. Resets at 2025-04-06T16:00:00Z"
}

The Retry-After header indicates the number of seconds to wait before retrying.

Monitoring Usage

Track your current usage on Studio's Usage page (/studio/settings/usage). It shows:

• Real-time quota bars for the 4-hour window and the monthly credit allowance
• Color-coded warnings (green < 80%, yellow 80-95%, red ≥ 95%)
• Historical request charts (7-day and 30-day views)

Upgrading Limits

If you consistently hit rate limits, upgrade your Token Plan's tier:

1. Go to Studio's Plan page (/studio/settings/plan)
2. Select a higher tier
3. New limits take effect immediately