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: Rigetti Computing
providerId: rigetti
created: '2026-05-25'
modified: '2026-05-25'
reconciled: false
tags:
  - Quantum Computing
  - QCS
  - Rate Limiting
  - QPU
description: Documented rate-limit and concurrency policies for the Rigetti QCS API surface. Rigetti does not
  publish a public per-second request budget for the QCS HTTP API; the operational rate-limiting comes from (1)
  reservation-based time-boxed exclusive QPU access, and (2) engagement scope on the gRPC Controller service.
  The HTTP API does enforce standard HTTP throttling responses (429) with a Retry-After header on abuse.
sources:
  - https://docs.api.qcs.rigetti.com/
  - https://docs.rigetti.com/qcs/guides/the-rigetti-qcs-api
  - https://docs.rigetti.com/qcs/troubleshooting
headers:
  retryAfter: retry-after
responseCodes:
  throttled: 429
  quotaExceeded: 429
algorithm: per-reservation-time-slice
limits:
  - tier: All QCS users
    surface: HTTP API (api.qcs.rigetti.com)
    rpm: not-documented
    notes: Account, billing, reservations, endpoints, and quantum-processor metadata endpoints. Token is valid
      24 hours after issuance; abusive callers receive 429 with retry-after.
  - tier: All QCS users
    surface: gRPC Controller service
    rpm: bounded-by-engagement
    notes: Execution is gated by engagement scope and QPU duty cycle. Only the active reservation holder can
      submit Controller jobs to a quantum-processor endpoint. BatchExecuteControllerJobs is the preferred
      high-throughput surface.
  - tier: All QCS users
    surface: gRPC Translation service
    rpm: not-documented
    notes: Quil-to-Controller-Job translation; pre-execution step. No documented public RPM.
quotas:
  - name: Reservation duration
    description: Reservations are billed for their duration regardless of utilization. Reservations are scoped
      to a specific quantumProcessorId.
    metric: minute
    enforcement: hard
  - name: Engagement TTL
    description: Engagements (gRPC access tokens) are short-lived and scoped to one endpoint; clients re-create
      them as needed.
    metric: token
    enforcement: hard
  - name: Access token TTL
    description: OAuth2 access tokens issued by Okta are valid 24 hours from issuance.
    metric: token
    enforcement: hard
notes: |
  Public QCS rate limits are not enumerated in the OpenAPI spec. Effective throughput on a Rigetti QPU is
  controlled by the reservation calendar (see /v1/quantumProcessors/{id}/calendar) and the engagement
  authorization mechanism rather than per-second request budgets. Indirect users via AWS Braket and Azure
  Quantum inherit those platforms' published rate limits.