generated: '2026-09-19'
method: searched
source: https://network.mercury-hq.com/pricing
sources:
- https://network.mercury-hq.com/pricing (tier table with a "Rate limit" column)
- https://network.mercury-hq.com/university/developers (key-issuance cap)
- https://network.mercury-hq.com/terms ("honours target robots/rate limits on a best-effort basis" - about upstream sites, not a limit on callers)
- response headers observed 2026-09-19 on GET /buy/fetch (402), GET /buy/signal (200), GET /buy/matter (503), POST /mcp tools/list (200), POST /a2a (200), GET /openapi.json (200)
- text search of openapi.json, llms.txt, /catalog and /.well-known/x402 for "rate limit", "429", "quota", "Retry-After"
limit_count: 4
description: >-
  Published limits exist only on the API-key rail and are per key per minute, stated in the /pricing tier
  table; the x402 rail publishes no request limit - an unpaid caller is stopped by the 402 payment
  challenge rather than a quota, and nothing on the docs, OpenAPI, llms.txt or discovery documents mentions
  a per-wallet ceiling. No rate-limit header (RateLimit-*, X-RateLimit-*) appeared on any probed response
  and the status code and body returned on exhaustion are undocumented, so an agent on the key rail has
  the number but no runtime signal. The one Retry-After observed (3600) was on the 503 mint-disabled
  response of a gated SKU, which is a delivery-off signal, not throttling.
limits:
- scope: per API key (Sandbox tier, mk_test_)
  window: 1 minute
  limit: 5
  burst: not stated
  quota: 100 credits per month
  source: https://network.mercury-hq.com/pricing
- scope: per API key (Pro tier, $29/mo)
  window: 1 minute
  limit: 120
  burst: not stated
  quota: 50,000 credits per month
  source: https://network.mercury-hq.com/pricing
- scope: per API key (Business tier, $199/mo)
  window: 1 minute
  limit: 600
  burst: not stated
  quota: 1,000,000 credits per month
  source: https://network.mercury-hq.com/pricing
- scope: per IP (sandbox key issuance, POST /api/dev/keys)
  window: 1 hour
  limit: 20 keys
  source: https://network.mercury-hq.com/university/developers ("issuance is capped at 20 keys per IP per hour")
- scope: per API key (Enterprise tier, $1500/mo)
  window: null
  limit: null
  note: the tier table shows "—" for rate limit and "unmetered" credits; no number is stated
- scope: x402 rail (keyless, per call)
  window: null
  limit: null
  note: no request limit published; each call is gated by a 402 challenge for the SKU price; per-call caps that do exist are payload caps (5s timeout, 10MB body, urls <=20 on /buy/batch, prevText/content <=256KB)
response_headers:
  observed: [retry-after (only on the 503 mint-disabled response, value 3600)]
  checked_for: [RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset, RateLimit-Policy, X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset, Retry-After]
  note: no rate-limit header on any 200, 402 or MCP response; headers seen were x-powered-by Express, etag, server Fly/..., via fly.io, fly-request-id
exhaustion_status: null
exhaustion_note: undocumented - neither 429 nor a credit-exhausted error shape is described; /developers says "Check your balance - GET /api/dev/balance" (401 without a key)
x-evidence:
  fetched: '2026-09-19'
  probes:
  - {url: 'https://network.mercury-hq.com/buy/fetch?url=https://example.com', method: 'GET (unpaid)', status: 402, rate_limit_headers: none}
  - {url: 'https://network.mercury-hq.com/buy/signal', method: GET, status: 200, rate_limit_headers: none}
  - {url: 'https://network.mercury-hq.com/buy/matter', method: GET, status: 503, headers: 'retry-after: 3600'}
  - {url: 'https://network.mercury-hq.com/mcp', method: 'POST tools/list', status: 200, rate_limit_headers: none}
  - {url: 'https://network.mercury-hq.com/api/dev/balance', method: GET, status: 401}
  - {url: 'https://network.mercury-hq.com/pricing', status: 200}