Skip to main content
Flash applies separate limits to its REST endpoints and its WebSocket connections.

​
REST

The Flash API enforces a rate limit of 5 requests per second on each endpoint, per API key: Each endpoint (quote, order, list orders, get order, update, cancel, batch cancel) has its own independent bucket, so hitting the limit on one does not throttle the others.
Need higher throughput? We can support custom limits per client based on your needs. Let us know the rate limits you’re looking for on each endpoint and we can set them up.

​
Response headers

Every response includes the following draft-6 rate-limit headers, scoped to the endpoint you called: On a 429 response, the same RateLimit-* headers are returned. Use RateLimit-Reset to know how many seconds until you can retry (at most 1, since the window is 1 second).
  • Read RateLimit-Remaining after each successful response and throttle when it nears zero.
  • On 429, wait the number of seconds in RateLimit-Reset before retrying.
  • Use exponential backoff with jitter for sustained pressure.

​
WebSocket

The Orders WebSocket is bounded by concurrent connections per API key rather than requests per second. The per-key connection limit is at least 50; if you need it raised, reach out. Each connection streams a single funder, so open a separate connection for every funder you track. See Connection for connection handling, heartbeats, and connection-level error codes (including CONNECTION_LIMIT).