Skip to main content
Rate limits cap how much work your organization can submit in a given time window, so one user or an accidental traffic spike cannot slow processing for everyone else. Limits apply per organization, so all API keys in an organization share them. When you exceed a rate limit, the API returns an HTTP 429 response. Wait before retrying and use exponential backoff.

​
Parse API

The Parse API has two kinds of limits:

​
Size Limits

Size limits apply per request and differ between synchronous requests and async jobs:

​
Rate Limits

Every request counts against one of two rate limits, one per service tier, and the two limits are independent of each other. Each parsing model family also has its own independent set of these limits, with the same values: pages parsed with DPT-3 Verity do not count against DPT-3 Pro’s limits, and vice versa. Synchronous requests and priority jobs count against the same priority limit, so splitting traffic across the two routes does not increase capacity. Jobs run on the standard tier by default. To set a job’s tier, see Choose a Service Tier.
The Explore values apply once you purchase credits. During the free trial, a lower standard limit applies.

​
Run Large Documents as Async Jobs on the Standard Tier

Synchronous requests are hard-capped at your plan’s per-minute limit. For async jobs, a priority job with more pages than your per-minute limit is accepted if the page count is within your rate limit’s larger allowance. A priority job exceeding that allowance returns a 429; the error message states the limit. Submit the document as a Parse Job on the standard tier, which accepts up to 6,000 pages per PDF.

​
Extract API

The Extract API has no input size limit, and rate limiting differs between synchronous requests and async jobs. The synchronous limit depends on your pricing plan. For jobs, the service tier affects turnaround time and credit consumption only. To set a job’s tier, see Choose a Service Tier.