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: Inkeep
providerId: inkeep
created: '2026-06-20'
modified: '2026-06-20'
reconciled: false
tags:
- AI
- Support
- RAG
- Agents
- Documentation
- Rate Limiting
- Quotas
- Throttling
description: >-
  Inkeep's AI / RAG chat completions endpoint applies per-IP rate throttling
  and bounds each chat session to roughly 30 messages, with recommended input
  of <=100 tokens and output of <=1,000 tokens per request. Inkeep does not
  publish exact numeric per-account RPM/TPM ceilings; effective limits depend
  on plan and quoted usage. Specific values are not reconciled in this artifact.
notes: >-
  Confirm per-endpoint and per-plan limits with Inkeep during reconciliation;
  the documented constraints below are the publicly stated ones.
sources:
- https://docs.inkeep.com/cloud/ai-api/chat-completions-api
- https://docs.inkeep.com/cloud/overview/developer-platform
- https://docs.inkeep.com/cloud/faqs/pricing
responseCodes:
  throttled: 429
limits:
- name: Per-IP Throttling
  scope: ip
  metric: requests
  limit: see provider documentation
  notes: The AI / RAG chat completions endpoint throttles requests per IP address.
- name: Messages Per Chat Session
  scope: session
  metric: messages
  limit: 30
  notes: A single chat session is bounded to approximately 30 messages.
- name: Recommended Input Tokens
  scope: request
  metric: tokens
  limit: 100
  notes: Recommended maximum input of about 100 tokens per request.
- name: Recommended Output Tokens
  scope: request
  metric: tokens
  limit: 1000
  notes: Recommended maximum output of about 1,000 tokens per request.
- name: Account Rate Limits
  scope: account
  metric: requests
  limit: see provider documentation
  notes: Per-account ceilings depend on plan and quoted usage; not publicly numbered.
policies:
- name: Plan-Based Limits
  description: Effective limits scale with the quoted self-serve or Enterprise plan.
- name: Backoff Strategy
  description: Clients should implement exponential backoff with jitter and honor 429 responses.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com