generated: '2026-09-19'
method: searched
docs: https://chilledsites.com/docs/api#rate-limits
limit_count: 4
summary: >-
  Rate limits are documented, with X-RateLimit-* headers and a 429 RATE_LIMIT_EXCEEDED envelope, but the
  provider publishes FOUR mutually inconsistent statements of the numbers across four pages. All four are
  recorded with their sources; an agent should trust the response headers over any of them.
headers:
- name: X-RateLimit-Limit
  example: '100'
- name: X-RateLimit-Remaining
  example: '95'
- name: X-RateLimit-Reset
  example: '1635174000'
  format: unix epoch seconds
- name: Retry-After
  documented: false
headers_note: 'Documented at /docs/api ("Every API response includes rate limit information in the headers"). Not observed: the only anonymous responses obtainable are 401s, which carried no X-RateLimit-* headers.'
exhaustion:
  status: 429
  body: '{"success":false,"error":{"code":"RATE_LIMIT_EXCEEDED","message":"Rate limit exceeded. Try again in 3600 seconds.","details":{"limit":100,"remaining":0,"resetAt":"2026-04-21T15:00:00Z"}}}'
  source: https://chilledsites.com/docs/api#rate-limits
limits:
- source: https://chilledsites.com/docs/api#rate-limits
  scope: per subscription tier, per hour
  window: 1 hour
  tiers:
    free: {website_generation: 10, read_operations: 1000, write_operations: 500, deploy: 50}
    pro: {website_generation: 100, read_operations: 10000, write_operations: 5000, deploy: 500}
    enterprise: {website_generation: unlimited, read_operations: unlimited, write_operations: unlimited, deploy: unlimited}
  note: The API reference table. Also states per-endpoint "Free Tier 10 generations/hour, Pro Tier 100 generations/hour" under POST /v1/generate. No Enterprise tier exists on the pricing page.
- source: https://chilledsites.com/agents.md
  scope: per key, per minute (also repeated in agent-meta.json rateLimits)
  window: 1 minute
  limits: {read_operations: 100, generation: 10, deploy: 5}
- source: https://chilledsites.com/developers
  scope: per API key
  window: 1 minute
  limits: {all_requests: 10}
  note: 'FAQ: "The API is rate-limited to 10 requests per minute per API key ... contact support for higher limits."'
- source: https://chilledsites.com/for-agents
  scope: per key / per account, per day
  window: 24 hours
  limits: {website_generations_per_key: 5, checkout_emails_per_account: 3}
  note: 'FAQ: "Five website generations per key per day, and three checkout emails per account per day. Both exist so a looping agent cannot burn your money or flood your inbox." The checkout-email cap is also stated under Guardrails as "a hard cap of three checkout emails per account per day".'
token_quotas:
  note: Separate from rate limits — every generation also debits a token balance (402 when short). See plans/ for the balances per tier and conventions/ for the 402 flow.
  documented_costs:
    agents_md: {website_generation: '~5,000 to start, typically 10,000-30,000 on completion', image: '~1,000', video: '~10,000', ad_copy: '~500', website_edit: '1,000-3,000', upload_update_deploy: 'no token charge, but paid account required'}
    docs_api: {website_generation: '500-1,000 tokens'}
    mcp_package_tool_descriptions: {generate_image: 'Gemini 2.5 Flash 1500 / Pro 2500 / Imagen 3 3500 / DALL-E 3 ...', generate_video: '8000 (Veo 3.1)'}
    note: The three sources disagree by an order of magnitude on website generation; recorded verbatim.