generated: '2026-09-02'
method: probed
source: https://pagesnap.142-93-197-141.sslip.io/docs#auth
probe:
  url: https://pagesnap.142-93-197-141.sslip.io/v1/meta?url=https://example.com
  http_status: 200
  observed: '2026-09-02'
  observed_headers:
    x-plan: anon
    x-ratelimit-limit: '30'
    x-ratelimit-remaining: '29'
    x-ratelimit-scope: day
    x-render-ms: '7'
  note: >-
    Headers below are not merely documented - they were read off a live unauthenticated response.
    X-RateLimit-Scope names WHICH window the counter belongs to, which is what lets an agent tell
    a daily cap from a monthly one without a second call.
response_headers:
  - name: X-Plan
    meaning: The plan the request was metered against (anon, free, starter, pro, scale).
  - name: X-RateLimit-Limit
    meaning: Ceiling for the scope named by X-RateLimit-Scope.
  - name: X-RateLimit-Remaining
    meaning: Requests left in that window.
  - name: X-RateLimit-Scope
    meaning: Which window the limit/remaining pair describes (e.g. day).
  - name: Retry-After
    meaning: Set on 429 and on 503 BUSY when a useful wait can be computed.
  - name: ETag
    meaning: Returned on cacheable content responses; 304 is supported.
exhaustion:
  status_codes:
    - code: 429
      codes: [RATE_LIMITED, TARGET_BUSY, ANON_GLOBAL_QUOTA, DAILY_QUOTA, MONTHLY_QUOTA, CONCURRENCY, KEY_LIMIT, KEY_SPIKE_LIMIT]
      meaning: Caller limit, per-target protection, or plan quota exhausted.
    - code: 503
      codes: [BUSY, ENGINE_UNAVAILABLE, JOB_QUEUE_FULL]
      meaning: >-
        Plan-prioritized admission wait expired, or sustained CPU load shed new anonymous browser
        work. Distinct from 429 on purpose - it is capacity protection, not a caller limit.
  guidance: Respect Retry-After and back off exponentially. Nothing is charged automatically on 429.
limit_count: 11
limits:
  - scope: per-ip
    plan: Anonymous
    window: day
    limit: 30
  - scope: per-ip
    plan: Anonymous
    window: minute
    limit: 10
  - scope: per-ip
    plan: Anonymous
    window: concurrent
    limit: 1
  - scope: per-key
    plan: Free
    window: month
    limit: 3000
    note: A 100/day cap applies simultaneously; both must be satisfied.
  - scope: per-key
    plan: Free
    window: day
    limit: 100
  - scope: per-key
    plan: Free
    window: minute
    limit: 20
    burst: 2 concurrent in-flight
  - scope: per-key
    plan: Starter
    window: 31-day term
    limit: 1500
    note: Pooled, no daily cap. 30 requests/minute, 2 concurrent.
  - scope: per-key
    plan: Pro
    window: 31-day term
    limit: 10000
    note: Pooled, no daily cap. 60 requests/minute, 3 concurrent.
  - scope: per-key
    plan: Scale
    window: 31-day term
    limit: 50000
    note: Pooled, no daily cap. 120 requests/minute, 5 concurrent.
  - scope: per-target-host
    plan: all
    window: minute
    limit: 60
    burst: 3 in flight
    error_code: TARGET_BUSY
    note: Protects each destination hostname Pagesnap fetches on the caller's behalf.
  - scope: per-key-and-per-ip
    plan: all
    window: day
    limit: 10 invoices per key, 20 per IP
    error_code: CHECKOUT_KEY_LIMIT / CHECKOUT_IP_LIMIT
    note: Checkout invoice creation. Failed provider attempts count toward the cap.
other_caps:
  - Batch size 3/10/15/25/50 URLs by plan.
  - Crawl effective limits over 25 pages return 202 and reuse the 24-hour job API.
  - llms-txt generation limit 1-200 pages.
  - Max 3 new API keys per IP per day.
  - Monitors 0/2/10/50/200 by plan; intervals 24h on Free, 1h/6h/24h on paid.
  - Monitor scheduler runs at most 2 checks concurrently with jitter.
change_policy: >-
  Free and anonymous limits may change prospectively for capacity or abuse control; the docs and
  response headers are updated when that happens. A paid plan's limits will not be reduced during
  its purchased 31-day service period. Limit increases and new overload safeguards are treated as
  additive changes.