Getting started
API Rate Limits
Default limits: execution planes 30 to 300 per minute by tier, models plane 600 per minute per key.
Every key has a tier, copied onto the key when it is created. The tier sets per-minute and per-day windows on the execution planes and the MCP server; the models plane has one per-key limit for every tier.
| Surface | Counted on | free | pro | enterprise |
|---|---|---|---|---|
Execution planes on https://api.plungeai.com (tools, agents, workflows, discovery, traces) | POST requests only; catalog GETs are not counted; a per-key override is ignored | 30/min · 1,000/day | 100/min · 10,000/day | 300/min · 100,000/day |
MCP server https://mcp.plungeai.com/v1 | tools/call only; a per-key rate_limit override set by PlungeAI is honoured | 30/min · 1,000/day | 100/min · 10,000/day | 300/min · 100,000/day |
Models plane (/v1/chat/completions, /v1/embeddings, /v1/models) | Every request, per key | 600 req/min | 600 req/min | 600 req/min |
| Failed authentication | Per client IP, any tier | 20/min · 500/day | n/a | n/a |
A key with no tier counts as free. When you hit a limit you get 429 with Retry-After (seconds): rate_limited on the execution planes, rate_limit_exceeded on the models plane, and JSON-RPC -32000 with error.data.retryAfter on MCP. There are no X-RateLimit-* headers; call plungeai_whoami to see your live window.
Pricing
Rate limits are separate from billing: insufficient_quota, spend_cap_exceeded and insufficient_balance come from billing caps, not from these windows. See Pricing.
Need higher limits?
Contact PlungeAI support and quote your x-request-id. Support contact: TBD by owner (TI-62).