specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: Rossum
providerId: rossum
created: '2026-05-08'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-08, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-08'
reconciled: true
tags:
- AI
- Document AI
- IDP
- Invoices
- OCR
- Rate Limiting
- Quotas
- Throttling
description: >-
  Rossum publishes a global REST API throttle of 10 requests per second on
  general endpoints, with 10 requests per minute specifically on the
  page-spatial-data translation endpoint. 429 with Retry-After is returned on
  excess. Documents up to 40 MB are accepted.
notes: >-
  Access tokens default to ~162 hours; long-lived tokens can be obtained via
  one-time-token exchange. Annotation processing is asynchronous (importing →
  to_review → exported) so polling cadence rarely needs to approach the per-
  second limit.
sources:
- https://elis.rossum.ai/api/docs/
responseCodes:
  throttled: 429
headers:
- name: Retry-After
  description: Seconds to wait before retrying after a 429.
limits:
- name: General REST endpoints
  scope: account
  metric: requests
  limit: 10
  timeFrame: 1s
- name: Page spatial data translation
  scope: account
  metric: requests
  limit: 10
  timeFrame: 1m
- name: Maximum upload size
  scope: request
  metric: MB
  limit: 40
  notes: Per-document upload size limit.
policies:
- name: Backoff Strategy
  description: Honor Retry-After on 429; use exponential backoff with jitter.
- name: Polling vs Hooks
  description: Prefer webhook (hook) extensions over polling annotations endpoints to stay below limits.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com