Daily quota

Each successful search takes 1 request from your daily quota.
  • The quota resets every day at 02:00, Paris time. quota.resets_at gives the exact moment in UTC.
  • It is shared with your dashboard on orelys.io: a search there and a search through the API take from the same quota.
  • Unused requests do not carry over to the next day.

What takes a request

Daily allowance. Searches that reach the archive but bring nothing back are free up to your daily quota each day — at least 10 — shared with your dashboard. Past that, they take a request like any other search, until the quota resets. A search service that is down, busy or too slow is always free. Every search response tells you: the charged field in JSON responses, and the X-Request-Charged header on all of them.

OSINT credits

The OSINT modules never touch the daily quota: they are paid in credits. Your balance: credits.balance in /v1/me, then credits.remaining and the X-Credits-Remaining header after each OSINT search.

Rate limits

How fast your account can search depends on your plan: These two limits belong to the account: searches with your API key and searches from your dashboard on orelys.io count together. A third limit applies to every address, whatever the plan.
  • The minute is a sliding window: a slot frees up 60 seconds after the call that used it.
  • Every search you send counts in the minute, even one refused for invalid input or because your searches at the same time are all in use. A search refused with rate_limited does not.
  • Your own numbers are in limits from /v1/me, and in the X-RateLimit-Limit header of every search.
  • A plan change applies from your next search.
  • The address limit counts searches too: from a single address, an Enterprise account’s 120 searches a minute use all of it.

Headers

Handling 429

Wait for Retry-After seconds, then try again. quota_exceeded is different: the daily quota is used up, and Retry-After points at the next reset — retrying sooner will not help.
Keep no more searches running than your plan allows — 3 on Professional, 5 on Enterprise (limits.concurrent_requests in /v1/me). The rest are refused rather than queued.