generated: '2026-06-20'
method: searched
source: https://docs.knock.app/concepts/environments
description: >-
  Knock's test-vs-live separation is environment-based rather than a single
  global "test mode." Every Knock account has isolated Environments (a
  Development environment plus Production, and optionally staging), and API keys
  are scoped per environment — so a Development key exercises the same API
  surface without touching production recipients or sending real messages
  through production channel configs. Knock does not publish magic test card
  numbers or fixed synthetic identifiers the way a payments provider does;
  testing is done against your own Development-environment data. Only values
  actually documented by Knock are recorded here — no test credentials are
  invented.
test_vs_live:
  model: per-environment API keys (Development vs Production), not a global test flag
  key_prefixes:
    public: pk_    # client-side; environment-scoped
    secret: sk_    # server-side; environment-scoped
  note: >-
    Key prefixes documented in the OpenAPI security scheme and auth docs are
    pk_ (public) and sk_ (secret); the environment (development/production) is
    encoded in the key issued from that environment's settings.
  docs: https://docs.knock.app/developer-tools/api-keys
sandbox_mode:
  available: true
  surface: Dashboard Test Runner
  capabilities:
  - Run workflows in sandbox mode without delivering real notifications.
  - Skip delay steps to validate the full execution path in seconds.
  source: https://knock.app/changelog  # "Test runner improvements", 2026-06-30
magic_test_values: []
test_clocks:
  available: false
  note: No published test-clock / time-simulation primitive; delay steps can be skipped in the Test Runner instead.
fixtures:
  note: Local resource fixtures are managed as files via the Knock CLI (knock pull/push) against a Development environment.