generated: '2026-08-10'
method: searched
source: >-
  https://amuncore.com/ (pricing table), https://amuncore.com/llms.txt,
  https://amuncore.com/openapi.json
provider: AmunCore
providerId: amuncore
description: >-
  AmunCore rate-limits by billing-period call quota rather than by requests per
  second. The quota is attached to the plan — a daily allowance on the free tier
  and a monthly allowance on the paid tiers — and exhausting it returns 429 with a
  Retry-After header. Per-application IP allowlisting and "rate limits" are also
  listed as platform features, but no per-second or burst figure is published.
sources:
- https://amuncore.com/
- https://amuncore.com/llms.txt
- https://amuncore.com/openapi.json
headers:
  limit: X-RateLimit-Limit
  remaining: X-RateLimit-Remaining
  reset: X-RateLimit-Reset
  retry_after: Retry-After
  note: >-
    The OpenAPI description states responses "may include" the three X-RateLimit-*
    headers; llms.txt states the 429 response includes Retry-After. None were
    observed on the unauthenticated probes, since those never reached a metered
    endpoint.
responseCodes:
  throttled: 429
  throttled_body:
    success: false
    error: RATE_LIMITED
    message: Monthly call quota exhausted
limit_count: 4
limits:
- name: Free plan call quota
  plan: Free
  scope: account
  metric: api_calls
  limit: 1000
  timeFrame: day
- name: Starter plan call quota
  plan: Starter
  scope: account
  metric: api_calls
  limit: 100000
  timeFrame: month
- name: Pro plan call quota
  plan: Pro
  scope: account
  metric: api_calls
  limit: 5000000
  timeFrame: month
- name: Enterprise plan call quota
  plan: Enterprise
  scope: account
  metric: api_calls
  limit: -1
  timeFrame: month
  note: Advertised as unlimited calls; governed by the negotiated agreement.
other_limits:
- name: Page size
  scope: request
  metric: records_per_page
  limit: 500
  source: openapi parameter description — "Records per page (max 500)"
introspection:
  tool: check_plan_limits
  transport: mcp
  note: >-
    The MCP surface exposes check_plan_limits and get_license_status, so an agent
    can read remaining quota before spending it rather than discovering exhaustion
    at the 429. That is a genuinely good agent affordance; there is no REST
    equivalent.
gaps:
- No per-second or burst limit is published, so a client cannot tell how fast it
  may drain its monthly quota.
- Quota is per account, not per application or per key, so one noisy consumer's
  key can exhaust the allowance for every other consumer of the same account.
- The X-RateLimit-* headers are documented as conditional ("may include"), which
  means an agent cannot rely on them for pacing.
evidence:
- url: https://amuncore.com/
  status: 200
  note: Pricing table lists 1K calls/day, 100K calls/mo, 5M calls/mo, unlimited.
- url: https://amuncore.com/openapi.json
  status: 200
  note: info.description documents the X-RateLimit-* headers and the 429 semantics.
- url: https://amuncore.com/llms.txt
  status: 200
  note: '"A 429 means the monthly call quota is used up; the response includes a Retry-After header."'