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: CMS Open Payments
providerId: open-payments
created: '2026-07-11'
modified: '2026-07-11'
reconciled: false
tags:
- Government Data
- Healthcare
- Open Data
- Rate Limiting
- Quotas
description: >-
  CMS Open Payments does not publish fixed numeric per-key rate limits, because
  the API requires no key or account. Practical throughput is governed by the
  page-size / limit parameters on each endpoint (search caps page-size at 100;
  datastore queries default to a 500-row limit and page via limit and offset) and
  by fair-use expectations for a shared public federal service. Very large pulls
  are better served by the CSV/JSON download endpoints or the bulk dataset files
  than by deep pagination of the query API.
notes: >-
  No documented per-minute or per-day request cap exists as of the review date.
  Consumers should paginate with limit/offset, prefer bulk download for whole-year
  extracts, honor any 429 or 503 responses with exponential backoff, and avoid
  hammering the shared endpoint. Numbers here are structural
  (page-size caps), not published throttles.
sources:
- https://openpaymentsdata.cms.gov/about/api
- https://dkan.readthedocs.io/en/latest/user-guide/guide_api.html
responseCodes:
  throttled: 429
limits:
- name: Datastore Query Row Limit
  scope: request
  metric: rows
  limit: 500 default per request
  notes: Set limit and offset to page through results; count in the response gives the full match size.
- name: Search Page Size
  scope: request
  metric: results
  limit: 100 max per page
  notes: The /search endpoint caps page-size at 100; use page to paginate.
- name: Public Request Rate
  scope: global
  metric: requests
  limit: not published
  notes: No fixed per-client request-rate limit is documented; fair use applies to the shared federal service.
policies:
- name: Fair Use
  description: A shared public federal API. Prefer bulk download endpoints for whole-dataset extracts rather than deep pagination.
- name: Backoff Strategy
  description: On 429 or 503 responses, back off exponentially with jitter before retrying.
- name: Pagination
  description: Use limit and offset on datastore queries and page / page-size on search and metastore listings.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com