generated: '2026-07-21'
method: searched
source: https://docs.thinknum.com/docs/getting-started.md
description: >-
  Cross-cutting request/response semantics for the Thinknum Data API, captured
  from the published documentation. REST over HTTPS to data.thinknum.com, JSON
  responses, token auth, a dated API-version header, offset-style pagination, and
  a fixed 20-requests-per-minute rate limit surfaced via HTTP 429. No idempotency
  key contract is documented.
base_url: https://data.thinknum.com
transport: https
media_type: application/json
authentication:
  style: bearer-token
  detail: authentication/thinknum-authentication.yml
versioning:
  style: header
  header: X-API-Version
  value: '20151130'
  detail: lifecycle/thinknum-lifecycle.yml
pagination:
  style: offset
  params:
    - name: start
      default: 1
      description: 1-indexed offset of the first row to return.
    - name: limit
      default: 100000
      description: Maximum number of rows to return.
  response_fields:
    - total
    - start
    - limit
  notes: >-
    Documented on the Company API (GET /companies/) list responses; `total`
    reports the full result count while `start`/`limit` window the page.
idempotency:
  supported: false
  notes: >-
    No Idempotency-Key header or idempotent-retry contract is documented. The
    504 Gateway Timeout guidance tells clients to retry, and retries attach to
    the existing queued query rather than starting a new one, but this is a
    query-queue behavior, not a request-idempotency guarantee.
rate_limiting:
  limit: 20
  window: minute
  throttled_status: 429
  detail: rate-limits/thinknum-rate-limits.yml
error_handling:
  style: http-status
  detail: errors/thinknum-problem-types.yml
  notes: >-
    Errors are signaled by HTTP status code (400/401/404/429/500/504). No
    application/problem+json envelope is documented.