generated: '2026-09-19'
method: searched
source: openapi/_original/moirailabs-com-openapi.json (POST /analytics/invocations 429 + Retry-After; WalletProfilingRateLimitResponse);
  https://moirailabs.com/a2a ("credit-aware rate limits"); live GET https://api.moirailabs.com/api/v1/plans/active
  2026-09-19 (no RateLimit-* headers returned)
limit_count: 5
summary: The only numbers Moirai Labs publishes are per-plan monthly quotas for "AI Chat with Agent" requests (5/10/50/200/300
  per month by tier) and per-plan contract slots. No requests-per-second/minute ceiling is published for the REST
  API, the A2A agent or the MCP bridge; the contract declares a 429 with Retry-After on one operation, the wallet-profiling
  responses carry an in-body {limit, remaining, resetAt} block, and the marketing copy says "credit-aware rate limits".
  An agent discovers the runtime ceiling only by hitting it.
rate_limits:
- name: AI chat with agent — Free
  scope: per-account (plan quota)
  window: 1 month
  limit: 5
  metric: AI chat requests
  burst: null
  source: https://moirailabs.com/pricing ("AI Chat with Agent" row of the plan comparison, compiled app constant)
  domains:
  - agent.moirailabs.com
- name: AI chat with agent — Enthusiast
  scope: per-account (plan quota)
  window: 1 month
  limit: 10
  metric: AI chat requests
  burst: null
  source: https://moirailabs.com/pricing ("AI Chat with Agent" row of the plan comparison, compiled app constant)
  domains:
  - agent.moirailabs.com
- name: AI chat with agent — Startup
  scope: per-account (plan quota)
  window: 1 month
  limit: 50
  metric: AI chat requests
  burst: null
  source: https://moirailabs.com/pricing ("AI Chat with Agent" row of the plan comparison, compiled app constant)
  domains:
  - agent.moirailabs.com
- name: AI chat with agent — Business
  scope: per-account (plan quota)
  window: 1 month
  limit: 200
  metric: AI chat requests
  burst: null
  source: https://moirailabs.com/pricing ("AI Chat with Agent" row of the plan comparison, compiled app constant)
  domains:
  - agent.moirailabs.com
- name: AI chat with agent — Enterprise
  scope: per-account (plan quota)
  window: 1 month
  limit: 300
  metric: AI chat requests
  burst: null
  source: https://moirailabs.com/pricing ("AI Chat with Agent" row of the plan comparison, compiled app constant)
  domains:
  - agent.moirailabs.com
signals:
  status_on_exhaustion: 429
  response_headers:
  - Retry-After (declared on 429 and 202 of POST /analytics/invocations; CORS-exposed)
  standard_headers:
    RateLimit-*: false
    X-RateLimit-*: false
    observed_on_live_200: none — the anonymous /plans/active response carried x-request-id and CORS headers only
  in_body:
    schema: WalletProfilingRateLimitResponse
    fields:
    - limit
    - remaining
    - resetAt (date-time)
    where: WalletProfilingResponse.rateLimit on POST /wallet-profiling/simple and /enriched
  quota_in_body:
    schema: WalletProfilingQuotaResponse
    fields:
    - limit
    - reserved
    - used
    - remaining
    - periodStart
    - periodEnd
    also: GET /subscriptions/current/wallet-profiling-quota
documented_scopes:
- scope: per-user
  applies_to: POST /analytics/invocations
  description: '"Per-user active invocation quota or create rate limit was exceeded" — two limits (concurrent active
    invocations, and create rate) behind one 429; neither number published'
- scope: per-plan (30-day period)
  applies_to: contracts under analysis; enriched wallet-profiling requests
  description: PlanResponse.contractAmount (1, 3, 5, 10 on the live plans) and enrichedProfilingRequestsPerPeriod
    (0 on every live plan) — plan quotas, not request rates
gaps:
- No requests-per-window number for any endpoint
- No RateLimit-*/X-RateLimit-* headers
- Synchronous analytics GETs declare no 429 at all
- The A2A/MCP host applies the same bearer gate; its limits are equally unpublished
note_on_scope: These are subscription quotas on the agent-chat surface, not HTTP rate limits; they are recorded
  because they are the only published, numeric usage ceilings and they bound what an autonomous agent on a given
  plan can ask.