Rate limits and quotas

Request rate

Limits are applied per IP address, per minute.

Endpoints	Limit
`/api/search/*`	60 req/min
`/api/billing/*`	60 req/min
`/api/auth/signin`, `/signup`, `/forgot-password`, `/reset-password`	10 req/min

Exceeding a limit returns 429:

{
  "success": false,
  "message": "Too many requests. Please slow down and try again.",
  "result": { "code": "rate_limited" }
}

Back off exponentially with jitter. Hammering a 429 extends it.

Rate limits are separate from credits. A 429 means you sent requests too quickly; a 402 means you ran out of balance.

Values above a maximum are clamped, not rejected. A limit of 500 returns 50 results and is billed for 50.

Field	Endpoint	Default	Maximum
`query` / `q`	`/api/search/query`	none	512 characters
`limit`	`/api/search/query`	10	50
`urls`	`/api/search/contents`	none	8 URLs
`days`	`/api/billing/usage`	30	365
`limit`	`/api/billing/usage`	100	500
`limit`	`/api/billing/ledger`	50	200
`limit`	`/api/billing/payments`	20	100
`name`	`/api/auth/keys`	`Untitled key`	60 characters

Batch your extractions

POST /api/search/contents takes up to 8 URLs in one call. Eight separate requests cost the same in credits but consume eight times the rate budget.

Re-fetch instead of re-searching

GET /api/search/searches/{id} returns a stored result. It is free and avoids spending both a charge and a rate slot on a query you already ran.

Skip enrichment when you only need links

enrich: false removes the per-page extraction cost and returns faster.