generated: '2026-08-11'
method: probed
source: https://weaveapi.dev/pricing/
docs: https://weaveapi.dev/docs/

limit_count: 0

summary: >-
  WeaveAPI publishes NO numeric rate limits. The pricing page's comparison table is the only place
  limits are mentioned at all, and it gives them as prose: "Standard account limits" for pay-as-you-go
  and "Account review required" for larger accounts. No requests-per-minute, tokens-per-minute,
  concurrency ceiling or burst allowance appears on the pricing page, the docs, the terms, or the abuse
  policy. Live unauthenticated responses were inspected on four routes and carry NO rate-limit response
  headers of any kind, so an agent has neither a documented number nor a runtime signal to back off on.

published_limits: []

documented_statements:
  - scope: account
    tier: Pay as you go
    text: Standard account limits
    source: https://weaveapi.dev/pricing/
    numeric: false
  - scope: account
    tier: Account review
    text: Account review required
    source: https://weaveapi.dev/pricing/
    numeric: false
  - scope: account
    text: >-
      "Rotating accounts, scraping keys, bypassing limits, or creating excessive load may result in
      immediate suspension." — limits are enforced as an abuse control, but never quantified.
    source: https://weaveapi.dev/abuse/
    numeric: false
  - scope: route
    text: >-
      "Model routes depend on upstream providers, network conditions, capacity limits, quota limits,
      and maintenance. We do not guarantee every route is available at all times."
    source: https://weaveapi.dev/terms/
    numeric: false

# What the wire actually returns — probed, not documented.
response_headers:
  method: probed
  fetched: '2026-08-11'
  rate_limit_headers_present: false
  checked_for:
    - X-RateLimit-Limit
    - X-RateLimit-Remaining
    - X-RateLimit-Reset
    - RateLimit-Limit
    - RateLimit-Remaining
    - RateLimit-Reset
    - RateLimit-Policy
    - Retry-After
  found: []
  headers_actually_returned:
    - name: x-oneapi-request-id
      example: '202608112040025861429178268d9d6T5viNnDh'
      note: Per-request correlation id. The docs ask for it when contacting support.
    - name: x-new-api-version
      example: v1.0.0-rc.10-abuse-hardening-20260731
      note: Gateway build identifier, returned on every response.
    - name: strict-transport-security
      example: max-age=31536000; includeSubDomains
    - name: x-content-type-options
      example: nosniff
    - name: via
      example: 1.1 Caddy
  probed_routes:
    - {method: GET, path: /v1/models, status: 401}
    - {method: POST, path: /v1/chat/completions, status: 401}
    - {method: GET, path: /v1/nonexistent, status: 404}
    - {method: GET, path: /api/pricing, status: 200}
  caveat: >-
    All four observations are unauthenticated. A rate-limit header set may be emitted only to
    authenticated callers or only once a limit is approached; this probe cannot rule that out. What it
    does establish is that no limit signal is available to a client before it has a key, and that the
    404/200 paths — which the gateway serves without auth — carry none either.

exhaustion_behavior:
  status_code: null
  note: >-
    Not documented and not observable without a key. The docs' troubleshooting table lists "Route
    temporarily unavailable" and "Insufficient balance" as failure modes and tells the developer to
    "try again later", but names no status code, no Retry-After, and no backoff guidance.

quota:
  model: prepaid-balance
  note: >-
    The real ceiling on a WeaveAPI account is the prepaid credit balance rather than a request rate.
    Balance exhaustion is surfaced in the docs as the "Insufficient balance" troubleshooting case.
  balance_visibility: Console only — no public or documented balance endpoint.
  balance_endpoints_probed:
    - {path: /v1/dashboard/billing/usage, method: GET, status: 401, note: Route exists (OpenAI-compatible billing path), undocumented by WeaveAPI.}
    - {path: /v1/dashboard/billing/subscription, method: GET, status: 401, note: Route exists, undocumented by WeaveAPI.}

recommendation: >-
  Publish a numeric per-key limit and emit RateLimit-Limit / RateLimit-Remaining / RateLimit-Reset (RFC
  9331 style) plus Retry-After on 429. A gateway whose value proposition is routing across providers is
  exactly the surface where an agent most needs to know when to back off and switch routes.

x-evidence:
  - url: https://weaveapi.dev/pricing/
    http_status: 200
  - url: https://api.weaveapi.dev/v1/models
    http_status: 401
  - url: https://api.weaveapi.dev/v1/chat/completions
    http_status: 401
  - url: https://weaveapi.dev/abuse/
    http_status: 200