Skip to content

Rate Limits

Rate limits are enforced per API key (not per account) on a sliding 60-second window:

Plan Requests / minute
Xplorer (free) 60
Builder 300
Scale 1,200

Each API key gets its own counter, so you can isolate workloads (e.g. one key per app or environment) without them competing for the same window.

When you exceed the limit, the API returns 429 Too Many Requests:

{
"error": "Rate limit exceeded",
"code": "RATE_LIMITED",
"limit_per_minute": 300,
"retry_after_seconds": 60
}

The Retry-After and X-RateLimit-Limit headers are also set.

Before your key is checked, every request also passes a per-IP guard of 1,200 requests per minute, for both the REST API and MCP. It stops abuse and key guessing; normal use never reaches it, so your plan limit above is the one that applies. If it trips you get a 429 with "code": "RATE_LIMITED" and a Retry-After header.

MCP tool calls use the same per-minute windows by plan, but on a separate counter from your REST traffic:

  • API-key sessions are counted per key (same as REST, separate bucket)
  • OAuth sessions are counted per user account

A rate-limited tool call returns an in-band tool error with "code": "RATE_LIMITED" and retry_after_seconds — no credits are deducted.

Beyond per-minute rate limits, each request costs credits. When your balance can’t cover a call, requests return 402 Payment Required:

{
"error": "Insufficient credits",
"code": "CREDITS_EXHAUSTED",
"cost": 3
}

Top up or upgrade on the billing page.

  • Use caching to reduce repeat calls for the same data
  • Space weather data updates every 1–5 minutes — polling faster wastes credits
  • Check the X-Pxie-Credits-Remaining response header to monitor your balance from code; the dashboard shows the full balance and expiry dates
  • Rate-limited requests (429) cost no credits — but back off using Retry-After