generated: '2026-08-23'
method: searched
source: https://docs.kore.ai/ai-for-service/apis/rate-limit
docs: https://docs.kore.ai/ai-for-service/apis/rate-limit
limit_count: 6
note: >-
  Kore.ai publishes a real rate-limit matrix for the AI for Service estate with named
  enforcement levels and numeric limits. What it does NOT publish is the runtime signal:
  no X-RateLimit-*/RateLimit-* response headers and no Retry-After are documented
  anywhere, and the docs describe exhaustion only as a human-readable error string. An
  agent therefore cannot see how close it is to a limit — it can only discover the limit
  by hitting it. The Agent Platform estate documents a 429 status and a RATE_LIMIT_EXCEEDED
  / QUEUE_FULL error code but publishes no numbers at all.
rate_limits:
- scope: per-bot
  product: Automation AI
  window: 60s and 1h (rolling, both enforced simultaneously)
  limit: 60 per minute
  secondary_limit: 1800 per hour
  applies_to: Bot Definition APIs and NLP Training APIs and Analytics APIs
  note: If the one-hour limit is breached, further requests are blocked even when the sixty-second budget is still available.
- scope: per-request-id
  product: Automation AI
  window: 60s and 1h (rolling)
  limit: 60 per minute
  secondary_limit: 1800 per hour
  applies_to: Bot Definition Status APIs and NLP Training Status APIs
- scope: per-account
  product: Automation AI
  window: 60s and 1h (rolling)
  limit: 60 per minute
  secondary_limit: 1800 per hour
  applies_to: Admin APIs
- scope: per-bot
  product: Automation AI
  window: 60s and 1h (rolling)
  limit: 600 per minute
  secondary_limit: 18000 per hour
  applies_to: Bot Definition Notification APIs (Proactive Notification, Proactive Notification Status)
- scope: per-account-per-table
  product: Automation AI
  window: 60s and 1h (rolling)
  limit: 600 per minute
  secondary_limit: 18000 per hour
  applies_to: Data / Data Table Access APIs (Insert, Insert Bulk, Update, Delete, Get Table Data, Get View Data)
- scope: per-account
  product: Contact Center AI and Quality AI
  window: 60s
  limit: 60 per minute
  secondary_limit: null
  applies_to: >-
    Contact Center AI analytics, agent, skills, queue, WFM, hours-of-operation, voice
    gateway and campaign APIs; Quality AI Raw Data and Auto QA Reporting APIs.
response_signalling:
  headers: []
  headers_note: No rate-limit response headers are documented on any Kore.ai surface.
  retry_after: false
  status_on_exhaustion: 429
  status_note: >-
    429 with code RATE_LIMIT_EXCEEDED is documented for the Agent Platform estate. The
    AI for Service docs give only the error strings "You've exceeded the per-minute rate
    limit. Please wait for sometime before retrying." and "You've exceeded the hourly rate
    limit. Please wait for sometime before retrying." without naming a status code.
  concurrency: >-
    The Agent Platform also returns 429 with code QUEUE_FULL when the execution queue is at
    capacity, and supplies a retryAfterMs value in the error payload rather than a header.
custom_limits:
  available: true
  how: Contact support for a higher or custom rolling rate limit.