generated: '2026-08-17'
method: searched
source: https://developers.kardinal.ai/reference/limits-and-quotas
limit_count: 0
published: false
note: >-
  Kardinal publishes a limits-and-quotas reference page, and that page states
  explicitly that rate limits, max payload size and SLA are "Not yet published" for
  BOTH sandbox and production, directing integrators to contact support@kardinal.ai
  to confirm the values provisioned for their account. This is a recorded, deliberate
  absence rather than a page we could not find — limit_count is 0 because no numeric
  limit is published anywhere, and no `RateLimits` pointer is wired into apis.yml
  because the provider does not publish a rate-limit signal.
response_headers:
  published: false
  x_ratelimit: false
  ratelimit_standard: false
  retry_after: false
  note: >-
    No X-RateLimit-*, RateLimit-* or Retry-After response header is documented in the
    docs or declared in the OpenAPI. An agent has no runtime signal to back off on.
limits: []
observed_429:
- operationId: requestNewMFAOTPCode
  path: POST /auth/mfa/resendOTP
  status: 429
  declared_in_spec: true
  limit: undocumented
  body: >-
    Returns 429 when called too often, but currently WITHOUT a structured JSON body —
    it does not follow the EnvelopedErrors format used by every other error.
- operationId: requestNewLoginOTPCode
  path: POST /login/resendOTP
  status: 429
  declared_in_spec: true
  limit: undocumented
  body: >-
    Returns 429 when called too often, but currently WITHOUT a structured JSON body —
    it does not follow the EnvelopedErrors format used by every other error.
throughput_constraint:
  name: simultaneous running plans
  field: status.waitingRoom
  published_value: null
  note: >-
    The one throughput ceiling the docs describe is account-specific and unpublished:
    when an agency already has its maximum number of simultaneously running plans, a
    new plan sits in a "waiting room" until a slot frees up. The threshold must be
    confirmed per account with support@kardinal.ai.
compute_budget:
  platform_cap: none
  field: maxOptimizationDuration
  note: >-
    There is no platform-wide cap on how long a plan may optimize; the integrator
    sets maxOptimizationDuration per plan.
quotas_by_environment:
  sandbox: {rate_limit: not published, max_payload: not published, sla: not published}
  production: {rate_limit: not published, max_payload: not published, sla: not published}
evidence:
- {url: 'https://developers.kardinal.ai/reference/limits-and-quotas', http_status: 200, fetched: '2026-08-17'}
- {url: 'https://developers.kardinal.ai/guides/sandbox-to-production', http_status: 200, fetched: '2026-08-17'}
- {url: 'https://developers.kardinal.ai/openapi.yaml', http_status: 200, fetched: '2026-08-17', note: two operations declare a 429 response}