generated: '2026-08-13'
method: searched
source: >-
  https://developer.dreamdata.io/server-side/server-side-tracking/ +
  https://developer.dreamdata.io/client-side/api/ +
  https://developer.dreamdata.io/mcp/mcp-server/ +
  https://dreamdata.io/pricing
checked: '2026-08-13'
limit_count: 0
limits: []
documented_constraints:
- surface: Event Tracking API
  endpoint: POST https://api.dreamdata.cloud/v1/batch
  kind: payload-size
  value: 500kb per request
  scope: per-request
  quote: "Dreamdata's endpoint supports sending batches of data, limited to 500kb per request."
  source: https://developer.dreamdata.io/server-side/server-side-tracking/
  note: >-
    A size ceiling, not a rate limit. The docs do not say what status code is returned when it
    is exceeded.
- surface: Platform
  kind: volume-entitlement
  value: monthly tracked users (MTUs) per plan
  scope: per-account
  source: https://dreamdata.io/pricing
  note: >-
    Billing-side volume metric, not a request-rate limit. The free plan additionally publishes
    Seats 5, Stage Models 3, Notifications 2, Syncs 1 — entitlements, not throttles.
response_headers:
  published: false
  observed: []
  note: >-
    No X-RateLimit-*, RateLimit-* or Retry-After headers are documented for any Dreamdata
    surface, and none were observable anonymously — the tracking endpoint requires a source API
    key and the MCP endpoint returns 401 before any rate-limit signalling.
status_on_exhaustion: not published
retry_behaviour:
  client_side: >-
    Both first-party SDKs "handle batching, retries, and graceful shutdown automatically"
    (Node.js and Go SDK pages), but the retry policy, backoff and the conditions that trigger a
    retry are not specified, so an agent cannot reason about them.
  source: https://developer.dreamdata.io/server-side/nodejs-sdk/
notes: >-
  Dreamdata publishes no rate limits for any surface: not for the batch tracking endpoint, not
  for the outbound webhook syncs, and not for the MCP server. An honest zero — the only
  documented request-level constraint is the 500kb batch ceiling. This is a real gap for agent
  callers, who have no published signal telling them when to back off.