generated: '2026-09-19'
method: searched
source: https://app.agentmesh.link/
corroboration: live unauthenticated responses from https://app.agentmesh.link (see probes)
checked: '2026-09-19'
summary: >-
  AgentMesh publishes exactly one limit and it is a plan quota, not a rate: the Pro tier includes
  "up to 100,000 API requests per month". No per-second or per-minute rate limit is documented
  anywhere, the OpenAPI declares no 429 response on any of its 48 operations, and no rate-limit
  header of any family was observed on live 200 or 401 responses. An agent therefore has no
  runtime pacing signal; the only way to know where it stands against the monthly cap is to call
  the usage/entitlement operations (see plans/).
limit_count: 1
scopes:
  - scope: per-account
    tier: Pro (CHF 19/month)
    surface: all API requests
    window: 1 month
    limit: 100000
    unit: requests
    burst: null
    status_on_exhaustion: not-published
    verbatim: "Jusqu'à 100'000 requêtes API par mois."
    source: https://app.agentmesh.link/
    note: >-
      A monthly quota attached to a paid tier. The Free and Business tiers publish no number at
      all, so a Free-plan agent — the tier machine self-registration lands on — cannot know its cap.
headers:
  documented: false
  observed: false
  response_headers: []
  absent:
    - X-RateLimit-Limit
    - X-RateLimit-Remaining
    - X-RateLimit-Reset
    - RateLimit
    - RateLimit-Policy
    - RateLimit-Limit
    - Retry-After
  observation_note: >-
    Observed live on 2026-09-19 without credentials: GET /health (200), GET
    /v1/agents/discover?capability=search&limit=3 (200), GET /v1/stats (401), GET /v1/agents/me
    (401), POST /a2a SendMessage (401). None carried any rate-limit or Retry-After header. The
    response header set is Cloudflare's (server, cf-ray, cf-cache-status, nel, report-to,
    alt-svc) plus the provider's own agentmesh-bootstrap and Link: rel="service-desc" headers.
statuses:
  rate_limited: not-published
  rate_limited_note: >-
    No 429 appears in the served OpenAPI (responses are 200/201/202 and 422 only) and no throttling
    behaviour is described in the docs, README, QUICKSTART, SKILL.md or llms.txt. Whether the
    monthly cap answers 429, 402 or 403 when exhausted is unobserved.
runtime_signal_alternative:
  operations:
    - {operationId: account_usage_v1_account_usage_get, path: 'GET /v1/account/usage'}
    - {operationId: m2m_entitlements_v1_m2m_entitlements_get, path: 'GET /v1/m2m/entitlements'}
    - {operationId: m2m_commercial_usage_v1_m2m_usage_get, path: 'GET /v1/m2m/usage'}
  note: >-
    These are the closest thing to a quota signal — pull, not push. Their response schemas are
    undeclared in the spec, and they answered 401 without a credential, so field names are unknown.
probes:
  - {url: 'https://app.agentmesh.link/health', status: 200, rate_limit_headers: none}
  - {url: 'https://app.agentmesh.link/v1/agents/discover?capability=search&limit=3', status: 200, rate_limit_headers: none}
  - {url: 'https://app.agentmesh.link/v1/stats', status: 401, rate_limit_headers: none}
  - {url: 'https://app.agentmesh.link/pricing', status: 404}