generated: '2026-08-26'
method: searched
source: https://docs.hyperstack.cloud/docs/api-reference/rate-limits
docs: https://docs.hyperstack.cloud/docs/api-reference/rate-limits
limit_count: 1
note: >-
  Hyperstack publishes exactly one documented ceiling and it is scoped to the SOURCE IP,
  not to the API key or the account. The docs are explicit that the limit applies across
  all endpoints from a single source IP, which means every agent, CI runner and human
  behind one NAT egress shares one budget. No RateLimit-* or X-RateLimit-* response
  headers are documented, so a client has no runtime signal - it can only observe the
  429 after it has already been throttled. Retry-After is not documented either.
rate_limits:
- scope: per-source-ip
  applies_to: all endpoints
  window: 1 minute
  limit: 500
  unit: requests
  burst: null
  exhaustion_status: 429
  exhaustion_behaviour: >-
    Every subsequent request from that IP returns 429 Too Many Requests for the remainder
    of the one-minute window.
  source: https://docs.hyperstack.cloud/docs/api-reference/rate-limits
response_headers:
  documented: false
  ratelimit_headers: []
  retry_after: false
  note: >-
    The rate-limit reference documents no response headers at all. Verified against the
    OpenAPI as well: no 429 response in openapi/nexgen-cloud-hyperstack-openapi.json
    declares a RateLimit-*, X-RateLimit-* or Retry-After header. This is the single
    largest agent-readiness gap in an otherwise strong API - an autonomous client cannot
    pace itself, only recover.
backoff_guidance:
  documented: true
  policy: exponential
  first_retry_wait_seconds: 1
  strategy: Double the wait on each subsequent retry; stop after 5 attempts and surface the failure.
  source: https://docs.hyperstack.cloud/docs/api-reference/rate-limits
escalation:
  available: true
  channel: support@hyperstack.cloud
  requirement: Use case, endpoints affected, and request volume.
interactions:
- topic: pagination
  note: >-
    The docs warn that request rate scales with page count when walking large result
    sets, and advise increasing page size rather than request rate above a few hundred
    items. See conventions/nexgen-cloud-conventions.yml.
ai_studio:
  documented: false
  note: >-
    The AI Studio API (console.hyperstack.cloud/ai/api/v1) publishes no separate rate
    limits. Whether the 500/min per-IP ceiling applies to it is not stated.