Rate limits and quotas

Requests are limited per API key. Exceeding a limit returns 429 with Retry-After in seconds.

Limits

Plan	Requests/minute	Concurrent async jobs	Uploads/minute
Developer (free)	60	2	20
Scale	300	10	20
Enterprise	custom	custom	custom
Test keys	20	2	10

Rate limits apply per key; credits pool per organization. A token bucket permits a 2× burst for up to 10 seconds, so short spikes succeed while sustained overload does not.

Headers

Text

X-RateLimit-Limit: 60
X-RateLimit-Remaining: 41
X-RateLimit-Reset: 1754827200

X-RateLimit-Reset is a Unix timestamp. It appears on every response, not only on 429s — read it and pace yourself rather than waiting to be throttled.

Concurrency is a separate budget

An in-flight async job holds a concurrency slot until it reaches a terminal state. On Developer, two running deep_review searches will make a third return 429 immediately. Slots are not queued.

What to do when you hit one

Honour Retry-After; do not retry immediately
Exponential backoff with full jitter, capped at 5 attempts
Cap client-side concurrency instead of racing the limiter — one worker pool sized to your plan beats N retrying threads
Spread bulk work: one extraction over 500 sources costs one slot; 500 single-source runs cost 500

Requesting an increase

Email api-support@scispace.com with your organization, the limit you need, expected sustained and peak volume, and a sample X-Request-Id. Typical turnaround is 2 business days.

Quotas are not rate limits

A quota caps total spend (credits) and returns 402. A rate limit caps throughput and returns 429. Handle them differently: 402 needs a top-up, 429 needs patience.

Changes to this policy

Limit reductions are announced 30 days ahead by email to organization owners and in the changelog. Increases take effect immediately.

rate-limit · errors-and-retries · pricing-and-credits · cost-control

Last updated 2026-08-10