Get startedRate limits & access

Rate limits & access

Access is by API key on a paid subscription, on one of two plans. Two rate limits apply to every key, and rate-limit headers on every response tell you exactly where you stand.

Plans & access

Access is by paid subscription — see the API access pageto subscribe. Both plans cover live prices and the full GPU & provider catalogue; the Pro plan adds historical price data and a higher quota.

PlanDaily quotaBurst limitPrice history
Developer250 / day60 / min
Pro1,000 / day120 / minIncluded
Two limits, both enforced. Requests are capped by a rolling per-minute burst and a daily quota (the numbers above, per plan). Hit either and you'll get a 429 until that window resets.
Price history is Pro-only. The price-history endpoint requires the Pro plan. A Developer key calling it gets a 403 with a FORBIDDEN error. Every other endpoint works on both plans.

Rate-limit headers

Every response carries your current standing, so you can throttle proactively rather than waiting to be rejected. The bare RateLimit-* headers describe the most constrained window; the per-window variants report the minute and day budgets explicitly.

HeaderDescription
RateLimit-LimitRequest ceiling for the most-constrained window.
RateLimit-RemainingRequests left in that window.
RateLimit-ResetSeconds until that window resets.
RateLimit-*-MinuteLimit / Remaining / Reset for the per-minute burst window.
RateLimit-*-DayLimit / Remaining / Reset for the daily quota window.
Retry-AfterOn a 429 only — seconds to wait before retrying.
Response headers
HTTP/1.1 200 OK
Content-Type: application/json
RateLimit-Limit: 120
RateLimit-Remaining: 117
RateLimit-Reset: 43
RateLimit-Limit-Minute: 120
RateLimit-Remaining-Minute: 117
RateLimit-Reset-Minute: 43
RateLimit-Limit-Day: 1000
RateLimit-Remaining-Day: 863
RateLimit-Reset-Day: 51840

When you hit a limit

A throttled request returns 429 with a RATE_LIMITED error and a Retry-After header. Back off for that many seconds, then retry — ideally with exponential backoff and jitter.

429 Too Many Requests
{
  "error": {
    "code": "RATE_LIMITED",
    "message": "Rate limit exceeded"
  }
}

Staying within limits

  • Request limit=200 on list endpoints to fetch more per call.
  • Cache reference data. The GPU and provider lists change rarely.
  • Watch RateLimit-Remainingand slow down before you're rejected.
  • On 429, always honour Retry-After instead of hammering.
© 2026 GPUs.io · API v1