generated: '2026-08-28'
method: searched
source: >-
  developer.specright.com — the full Specright API v1.1.0 reference (all 46 operations and
  their response tables), plus https://www.specright.com/plans-pricing/ and the 2020 API
  launch post. Searched 2026-08-28.
docs: https://developer.specright.com/api-reference
name: Specright API rate limits
description: >-
  Specright publishes no rate limits. A full-text search of the API reference for "rate
  limit", "429", "Retry-After", "quota" and "throttle" returned zero matches; no operation
  documents a 429 response; and no rate-limit response header appears anywhere. The pricing
  page attaches no call allowance to any tier. This is a complete absence of both the
  documented number and the runtime signal.
limit_count: 0
limits: []
headers:
  request: []
  response: []
  documented: false
  note: >-
    No X-RateLimit-*, no RateLimit-* (RFC 9331 style), and no Retry-After. An agent has no
    way to learn its remaining budget, and no way to learn how long to wait after being
    throttled.
exhaustion:
  status_code: not documented
  retry_after: not documented
  note: >-
    429 does not appear on any of the 46 operations. The documented error statuses are 400,
    401, 403, 406 and 422 only — see ../errors/specright-error-codes.yml.
burst: not documented
concurrency: not documented
quota_by_plan: not documented
bulk_guidance:
  note: >-
    The API's answer to volume is the asynchronous bulk-job family (CSV upload, 202 Accepted,
    poll for status) rather than a documented per-request rate. But no file-size ceiling, row
    ceiling, concurrent-job limit or polling interval is published either, so the bulk path
    is as unbounded on paper as the synchronous one.
underlying_platform_note: >-
  Specright runs on Salesforce, which enforces its own org-level API request limits. Those
  limits are real and will be hit, but they are Salesforce's and are not surfaced, mapped or
  even mentioned in Specright's reference — so a consumer cannot tell whether a failure is a
  Specright limit, a Salesforce org limit, or neither. This is recorded as context, not as a
  Specright-published limit.
findings:
  - >-
    Zero published limits and zero runtime signalling. An agent cannot pace itself, cannot
    detect throttling from a header, and cannot back off from a Retry-After.
  - >-
    Because pricing is quote-only, there is no plan-level allowance to infer a limit from
    either.
  - >-
    Bulk endpoints exist but publish no size or concurrency ceiling.