generated: '2026-08-13'
method: searched
source: https://docs.interchange.io/v2/reference/rate-limits
docs: https://docs.interchange.io/v2/reference/rate-limits
applies_to: Scope3 Interchange v2 Buyer and Storefront APIs (https://api.interchange.io/api/v2/*)
limit_count: 4
exhaustion:
  status: 429
  error_code: RATE_LIMITED
  envelope: '{"data": null, "error": {"code": "RATE_LIMITED", "message": "Too many requests"}}'
headers:
  standard: IETF RateLimit header fields, draft-7
  returned:
  - name: Retry-After
    meaning: Seconds to wait before retrying. Scope3 documents this as always to be honored.
  - name: RateLimit-Limit
    meaning: Max requests allowed in the current window
  - name: RateLimit-Remaining
    meaning: Requests remaining in the current window
  - name: RateLimit-Reset
    meaning: Seconds until the window resets
  guidance: >-
    Scope3 documents reading RateLimit-Remaining on every response, not only on 429s, and
    slowing down preemptively when it is low.
limits:
- scope: per-ip
  surface: General API endpoints
  window: 1 minute
  limit: 100
  unit: requests
- scope: per-ip
  surface: Auth endpoints (/auth/*)
  window: 15 minutes
  limit: 40
  unit: requests
  environment: production
- scope: per-ip
  surface: Signup endpoints
  window: 1 hour
  limit: 20
  unit: requests
  environment: production
- scope: per-ip
  surface: Password reset
  window: 1 hour
  limit: 12
  unit: requests
  environment: production
per_account:
  documented: true
  values_published: false
  note: >-
    Scope3 states per-account quotas are tier-based and may differ from the per-IP defaults,
    and directs customers hitting RATE_LIMITED consistently to support@scope3.com for a
    higher-tier quota. The tier values themselves are not published.
polling_cadences:
  note: Recommended cadences for long-running async endpoints, published alongside the limits.
  entries:
  - endpoint: GET /campaigns/:id/media-buy-status
    cadence: every 30s minimum
    note: Media-buy state changes are minute-scale.
  - endpoint: GET /discovery/:id/discover-products
    cadence: every 5-10s while pending
    note: Back off to 15s after one minute of pending state; slower for large agent sets.
  - endpoint: GET /campaigns/:id after POST /execute
    cadence: every 15-30s
    note: Scope3 recommends webhooks where available to remove the need to poll.
measurement_apis:
  documented: false
  note: >-
    The Carbon Calculator (api.scope3.com/v2) and AI Impact Measurement (aiapi.scope3.com)
    APIs publish no explicit throttling; the published guidance is to batch measurement rows
    (4,000-8,000 rows per request; requests over ~100,000 rows commonly fail) rather than a
    fixed request-rate ceiling. See conventions/scope3-conventions.yml.
cross_links:
  errors: errors/scope3-problem-types.yml
  conventions: conventions/scope3-conventions.yml