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: Crunchbase
providerId: crunchbase-data
created: '2026-07-11'
modified: '2026-08-14'
generated: '2026-08-14'
method: searched
source: https://data.crunchbase.com/docs/using-the-api
reconciled: false
tags:
- Company Data
- Funding Data
- Web Intelligence
- Rate Limiting
- Quotas
description: >-
  Crunchbase documents a single API-wide rate limit of 200 calls per minute across the
  Data API endpoints, applied to protect platform reliability and performance. Beyond
  that per-minute cap, usage is governed by the terms of the account's licence
  (Enterprise, Applications, or Basic) rather than by published per-endpoint quotas.
  Response sizes are separately bounded: search returns 50 items by default and 1000
  maximum, autocomplete caps at 25, and a card included inline on an entity lookup
  returns at most 100 items. Keyset cursors (after_id / before_id) traverse anything
  larger.
notes: >-
  The material gap is on the response side, not the request side. Crunchbase publishes
  NO rate-limit response headers - no RateLimit-Limit / RateLimit-Remaining /
  RateLimit-Reset, no X-RateLimit-*, and no Retry-After on 429. The documented remedy
  for repeatedly hitting the limit is "reach out to your CSM". An agent therefore has
  no runtime signal for remaining quota and must budget entirely client-side. Total
  monthly call allowances and any per-licence quotas are contractual and unpublished.
  An API Dashboard with usage trends and endpoint-level visibility was announced for
  Enterprise customers in June 2026, but it is a UI surface - there is no metering REST
  endpoint. The MCP server exposes cb_metering_get_state for the same need.
sources:
- https://data.crunchbase.com/docs/using-the-api
- https://data.crunchbase.com/docs/using-search-apis
- https://data.crunchbase.com/docs/using-autocomplete-api
- https://data.crunchbase.com/docs/getting-more-results-from-card_id
- https://about.crunchbase.com/product-updates/
responseCodes:
  throttled: 429
  httpsRequired: 426
  unauthorized: 401
responseHeaders: []
responseHeadersNote: >-
  Verified absent from the documentation and from all six published OpenAPI documents -
  no operation declares a rate-limit response header, and 429 is not declared as a
  response at all. Recorded as an honest zero.
limit_count: 5
limits:
- name: API Calls
  scope: account
  metric: requests
  limit: 200 per minute
  window: 60s
  notes: Documented API-wide rate limit across all Data API endpoints, per API key.
  source: https://data.crunchbase.com/docs/using-the-api
- name: Search Result Size
  scope: request
  metric: items
  limit: 50 default / 1000 max
  notes: >-
    Per-request result cap on POST /data/searches/{collection}; page with after_id /
    before_id. The real ceiling within 1000 depends on the licence.
  source: https://data.crunchbase.com/docs/using-search-apis
- name: Autocomplete Result Size
  scope: request
  metric: items
  limit: 10 default / 25 max
  notes: The limit query parameter on GET /data/autocompletes maxes at 25. No pagination.
  source: https://data.crunchbase.com/docs/using-autocomplete-api
- name: Entity Card Inline Items
  scope: request
  metric: items
  limit: 100 per card
  notes: >-
    Cards included via card_ids on an entity lookup return at most 100 items; use
    GET /data/entities/{collection}/{entity_id}/cards/{card_id} with a keyset cursor
    for more.
  source: https://data.crunchbase.com/docs/getting-more-results-from-card_id
- name: Licence Call Allowance
  scope: account
  metric: requests
  limit: per contract
  notes: >-
    Total call volume is governed by the Enterprise / Applications / Basic licence
    terms; not publicly published.
policies:
- name: HTTPS Required
  description: All API requests must use HTTPS; non-HTTPS requests are rejected with HTTP 426.
- name: Keyset Pagination
  description: >-
    Search, card and deleted-entity endpoints use keyset cursors (after_id / before_id
    over entity UUIDs) so results stay stable as the graph changes.
- name: Backoff Strategy
  description: >-
    Clients should stay under 200 calls/minute and implement exponential backoff with
    jitter on 429. Because no Retry-After is returned, the backoff schedule has to be
    chosen by the client rather than honoured from the response.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com