generated: '2026-08-25'
method: searched
source: https://mgemi.com/llms.txt
docs: https://mgemi.com/agents.md
limit_count: 0
note: >-
  M.Gemi documents that rate limiting exists but publishes no number. The store's own
  agent instructions say "Respect rate limits. The MCP endpoint is rate-limited per IP.
  Back off on 429 responses." No limit, window, burst or quota is stated anywhere on the
  store, and no RateLimit-* / X-RateLimit-* / Retry-After header was returned on any
  observed response. An honest zero: the signal is qualitative only.
limits: []
exhaustion:
  status_code: 429
  documented: true
  source: https://mgemi.com/llms.txt
  scope: per IP
  guidance: Back off on 429 responses.
response_headers:
  rate_limit_headers_returned: false
  probed: 2026-08-25
  checked_for: [RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset, X-RateLimit-Limit,
    X-RateLimit-Remaining, Retry-After]
  observed_instead:
  - name: shopify-complexity-score
    example: '1350'
    detail: >-
      Returned on the MCP tools/list response. A Shopify platform cost-of-request signal,
      not a documented rate-limit budget - M.Gemi does not document it and no
      remaining/limit companion header is returned, so an agent cannot compute headroom
      from it.
  - name: shopify-complexity-score-v2
    example: '135'
  - name: x-request-id
    detail: Correlation id on every response.