generated: '2026-08-13'
method: searched
source: https://apidocs.nextroll.com/guides/get-started.html
note: >-
  AdRoll has no test mode. There is no sandbox account, no test-vs-live key
  prefix, no magic test identifiers and no fixture tooling — every credential
  addresses production data and production ad spend. What DOES exist is an
  interactive query console and one server-side validation flag, both captured
  below. Nothing here is invented; the absences are recorded as absences.

test_mode:
  exists: false
  key_prefixes: null
  test_credentials: null
  magic_values: null
  test_clock: null
  note: >-
    The nearest published guidance is organisational, not technical: "It is
    recommended that you create an application for production and one for
    testing. This allows you to have a separate OAuth 2.0 redirect URI for each
    of your development environments." Two application registrations against the
    same live account — separate credentials, identical blast radius.

console:
  name: GraphiQL — GraphQL Reporting API console
  url: https://app.adroll.com/reporting/graphiql
  kind: interactive query console
  auth: signed-in AdRoll dashboard session
  read_only: true
  description: >-
    NextRoll points developers here to browse the reporting schema and build
    queries interactively: "You can experiment with building queries using
    GraphiQL." The Reporting API is read-only, so the console cannot mutate
    campaigns — which makes it the one genuinely safe place to explore this API.
  source: https://apidocs.nextroll.com/graphql-reporting-api/overview.html

dry_run:
  surface: Server-to-Server (S2S) Event API
  endpoint: 'POST https://srv.adroll.com/api?advertisable=<ADVERTISABLE_EID>'
  parameter: dry_run
  value: 'true'
  behaviour: >-
    "When used, the payload is validated and logged, but doesn't impact your
    audiences or attribution." A payload-validation mode for event ingestion —
    the only non-destructive write path AdRoll documents.
  source: https://apidocs.nextroll.com/server-to-server-api/reference.html

draft_first:
  surface: AdRoll MCP Server
  behaviour: >-
    Campaigns created through the MCP server are staged as drafts for human
    review before going live. A safety control on agent writes, not a sandbox —
    the draft is created in the real account.
  source: https://www.adroll.com/blog/adroll-mcp-server

campaign_draft_status:
  status: draft
  meaning: 'The campaign is in draft state and has not yet been launched.'
  note: >-
    The CRUD API's campaign status vocabulary includes `draft`, so a campaign can
    be created and inspected in production without serving impressions. This is
    the practical substitute for a sandbox on the write path.
  source: https://apidocs.nextroll.com/crud-api/api-usage.html

verification_tools:
- {name: PixelPal, purpose: 'verify the pixel is firing', url: 'https://help.adroll.com/hc/en-us/articles/360040010352'}
- {name: Pixel Assistant, purpose: 'verify pixel implementation', url: 'https://help.adroll.com/hc/en-us/articles/8261197025165'}

gap: >-
  An agent or an integrator cannot exercise campaign creation, ad upload,
  audience mutation or billing paths without touching a live advertising account.
  For a platform whose write operations spend money, the absence of a test mode
  is the single largest onboarding risk on this API.

x-evidence:
  checked: '2026-08-13'
  probes:
  - {url: 'https://apidocs.nextroll.com/graphql-reporting-api/overview.html', http_status: 200}
  - {url: 'https://apidocs.nextroll.com/server-to-server-api/reference.html', http_status: 200}
  - {url: 'https://apidocs.nextroll.com/guides/get-started.html', http_status: 200}