generated: '2026-09-19'
method: searched
source: https://dynamicfeed.ai/docs and /connect (access-policy table), https://dynamicfeed.ai/llms.txt, https://dynamicfeed.ai/info
  (free_tier), https://dynamicfeed.ai/dashboard, operation descriptions for /v1/batch, /v1/pro/batch and /v1/stream;
  headers probed live 2026-09-19 on 200/401/402 responses
docs: https://dynamicfeed.ai/docs
limit_count: 7
summary: Limits are expressed as request-shape caps (calls per batch, SSE bounds) and per-key daily quotas rather
  than requests-per-second windows. The keyless MCP endpoint is explicitly unmetered fair-use. No rate-limit response
  headers are emitted, and the exhaustion status code is not published.
rate_limits:
- name: Keyless MCP
  scope: per-endpoint
  applies_to:
  - https://dynamicfeed.ai/mcp
  - https://dynamicfeed.ai/sse
  limit: null
  window: null
  metric: request
  policy: fair-use, "No per-key quota; subject to service availability"
  source: https://dynamicfeed.ai/docs
- name: Keyless REST batch
  scope: per-request
  applies_to:
  - POST /v1/batch
  limit: 20
  window: request
  metric: tool calls per request
  policy: fair-use; "Testing and low-volume read-only batch access"
  source: https://dynamicfeed.ai/docs
- name: Batch per-call deadline
  scope: per-call
  applies_to:
  - POST /v1/batch
  limit: 20
  window: seconds
  metric: upstream deadline per tool call
  source: openapi operation description v1_batch_v1_batch_post
- name: Paid batch
  scope: per-request
  applies_to:
  - POST /v1/pro/batch
  limit: 100
  window: request
  metric: tool calls per request
  price: 0.001 USDC per call
  source: https://dynamicfeed.ai/v1/pro/catalog
- name: SSE stream bounds
  scope: per-connection
  applies_to:
  - GET /v1/stream
  limit: 60
  window: connection
  metric: updates (count 1-60); interval 5-300 s
  source: openapi operation description v1_stream_v1_stream_get
- name: Free key daily quota
  scope: per-key
  applies_to:
  - direct feed GETs and per-user routes with X-API-Key
  limit: 1000
  window: day (resets midnight UTC)
  metric: request
  conflict: llms.txt REST section and /dashboard say 100 requests/day; /info free_tier and llms.txt pricing say
    1,000/day. Recorded as 1,000 with the conflict flagged.
  source: https://dynamicfeed.ai/info
- name: Paid key daily quotas (llms.txt tiers)
  scope: per-key
  applies_to:
  - direct REST with X-API-Key
  limit: 25,000 / 250,000 / 10,000,000
  window: day
  metric: request
  tiers:
  - Pro
  - Team
  - Enterprise
  source: https://dynamicfeed.ai/llms.txt
  note: The current /pricing page no longer states quotas ("No quotas … are promised on this page").
response_headers:
  declared: []
  observed: []
  probe: HTTP 200 on POST /v1/batch, 401 on GET /status and 402 on POST /v1/pro/tool/current_time carried no RateLimit,
    X-RateLimit-* or Retry-After header (2026-09-19).
  exhaustion_status: 'undocumented (dashboard text: "DAILY QUOTA REACHED … resets at midnight UTC")'
quota_introspection:
  endpoint: GET /v1/me
  auth: X-API-Key
  returns:
  - plan
  - calls used today
  - daily quota
  - remaining
  note: The runtime signal exists but only as a separate keyed call, not on the throttled response itself.
gaps:
- No per-second/minute rate window published for any surface.
- No rate-limit headers.
- No documented 429 behaviour or Retry-After.
- Free-tier quota stated inconsistently (100 vs 1,000/day).