Concepts

Rate limits

Per-key, per-organization and per-source request limits, the headers that report them and how to back off.

Limit	Default
Per key: reads	120 a minute
Per key: writes	60 a minute
Per organization (all keys and partners; live and test counted separately): reads	300 a minute
Per organization: writes	120 a minute
Per partner key	600 a minute in total, plus each organization's limits
Per inbound source	60 a minute by default, plus a daily limit (default 500; at most 2,000, or 500 in URL mode)
Mockup generations	3 running at once per organization
Test webhooks	30 test sends an hour, 500 test deliveries a day per organization

Over a limit the answer is 429 rate_limited (or daily_cap_reached) with Retry-After in seconds. A 429 for a per-minute limit also carries RateLimit-Policy, RateLimit and X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset.

Staying under#

Use webhooks or GET /v1/events?sort=sync_asc instead of polling every list.
Use limit=100 for syncs and follow next_cursor.
Back off on 429: wait Retry-After, then retry with the same Idempotency-Key.

Need more? E-mail support@light-ai.app with your use case. The per-minute key and organization limits are the same for every organization; account caps (the number of keys, inbound sources and webhook endpoints, and the daily lead cap) can be raised per organization.