generated: '2026-08-23'
method: searched
source: https://iqm.tech/products/iqm-resonance/
limit_count: 0
note: >-
  HONEST ZERO. IQM publishes no numeric rate limit anywhere public — not in the Resonance
  product page, not in the docs at docs.iqm.tech, and not in the Apache-2.0 iqm-client
  source. What can be established is that limiting EXISTS: the first-party error taxonomy in
  iqm-station-control-client maps HTTP 429 to a dedicated RateLimitError class, so the server
  does throttle. It just never says at what threshold, over what window, or with what
  Retry-After contract. An agent cannot plan against this; it can only react to a 429.
  Separately, the real scarcity control on this API is not requests-per-second at all — it is
  the credit quota and the job queue, which are recorded in plans/.
limits: []
response_headers:
  observed: []
  documented: []
  note: >-
    No X-RateLimit-*, RateLimit-* or Retry-After header is documented, and none could be
    observed anonymously because every path on resonance.iqm.tech returns 401 before any
    limit is evaluated.
status_on_exhaustion: 429
error_class: RateLimitError
error_source: >-
  https://pypi.org/project/iqm-station-control-client/ —
  src/iqm/station_control/interface/errors.py (_ERROR_TO_STATUS_CODE_MAPPING)
quotas:
- scope: per-account
  tier: Starter
  limit: 30 credits per month
  unit: credits (QPU seconds)
  conditional: "based on availability"
  source: https://iqm.tech/products/iqm-resonance/
- scope: per-account
  tier: Pay-as-you-go
  limit: null
  unit: prepaid credits that do not expire
  note: No cap; execution is serialized through a first-come, first-served queue.
- scope: per-account
  tier: Skip-the-queue
  limit: null
  unit: prepaid credits plus reserved timeslots
  note: Queue bypass via guaranteed exclusive access during reserved timeslots.
client_side_controls:
- {name: IQM_CLIENT_REQUESTS_TIMEOUT, default: 120, unit: seconds, description: Per-request HTTP timeout enforced by iqm-client.}
- {name: IQM_CLIENT_SECONDS_BETWEEN_CALLS, default: 1.0, unit: seconds, description: Interval iqm-client waits between job-status polls — a client-side politeness control, not a server limit.}