generated: '2026-08-31'
method: searched
source: https://accuracite.com/api-docs
name: AccuraCite API Rate Limits
docs: https://accuracite.com/api-docs
limit_count: 2
limits:
- scope: per-key
  endpoint: POST /api/v1/verify
  window: 1m
  limit: 60
  unit: requests
  burst: null
  note: 'Docs: "Verification: 60 requests per minute".'
- scope: per-key
  endpoint: POST /api/v1/generate
  window: 1m
  limit: 30
  unit: requests
  burst: null
  note: 'Docs: "Generation: 30 requests per minute".'
quotas:
- scope: per-account
  window: monthly
  unit: credits
  note: >-
    Verification and generation draw on ONE shared monthly credit pool, not separate
    allowances. A successful verify deducts 1 credit; a generate deducts 4 credits per
    citation returned. Plan totals: Free 50/mo, Pro 1,000/mo, Bulk Verification
    10,000/mo. See plans/accuracite-plans-pricing.yml.
  source: https://accuracite.com/pricing
batch_semantics:
  endpoint: POST /api/v1/verify
  max_items: 10
  all_or_nothing: true
  note: >-
    A `texts` batch claims credit for the whole batch up front. If the remaining
    balance cannot cover every citation, the entire request is rejected with 429 and
    nothing is charged or processed — the API never partially verifies a batch.
exhaustion:
  status: 429
  meaning: >-
    429 covers BOTH the per-minute rate limit and the credit balance running out; the
    docs do not distinguish the two in the status code, so a client cannot tell a
    retry-after-a-minute condition from a top-up-your-account condition from the
    status alone.
response_headers:
  documented: []
  observed: []
  note: >-
    No X-RateLimit-*, RateLimit-* or Retry-After header is documented, and none was
    present on the live 401 observed on 2026-08-31 (probed:
    POST https://accuracite.com/api/v1/verify with no key). The limits are published
    as prose only, so an agent has no runtime signal of how much budget is left and
    must infer exhaustion from a 429 after the fact. This is the single clearest
    runtime-semantics gap on this API.