Setup, seed & teardown

Most tests need a starting state — a logged-in account, a record to act on, a clean environment — and some need cleanup afterward. Cofactor handles this with three flow lifecycle roles and two nodes that pull prepared data into a flow.

01Flow lifecycle roles

Every flow has a type. Standard flows are your actual tests. The other three run around them:

You set the role on the flow itself. When a suite runs, Cofactor runs seeds and setup first, then your standard tests, then teardown — so each test starts from a known state and the environment is left clean.

Seed flows run on a suite regardless of their promotion state, so the data your tests rely on is always provisioned first.

02Seeding data into a test

A seed flow provisions data — for example, creating an account or an order — and exposes the values your test needs. To make a value available downstream, the seed flow must declare it as an output (from an action that captures it, or from credential outputs).

In the test that needs that data, add a Seed node that references the seed flow. Its declared outputs become variables in your flow:

$var.{seed_name}.{output_key}

For example, a seed that creates a client and outputs id makes $var.create_client.id available to later steps. See Variables & macros for how to reference and reuse these values.

A Seed node can:

• Pin a version of the seed flow so your test always runs against the same provisioning logic.
• Choose what a seed failure does — block the test (treat missing data as a failure) or warn and continue.

03Running a prerequisite flow inline

A Setup node references another published flow to run as a prerequisite of the current flow — useful for shared data creation or environment preparation you don't want to duplicate across tests. Use a Setup node when one flow should always run another first; use the Setup flow role when the preparation applies to a whole suite.

04Managing seed data from the catalog

The seed-data catalog is where you review what a seed provisions and manage cleanup:

• Inspect provisioned records — open a seed entry to confirm setup completed and see the entities it created.
• Edit declared outputs — rename or expose the values downstream steps reference; keep names specific. If a seed provisions a login, declare the fields (email, password) so a Login node can map them.
• Add or toggle teardown — a Teardown button appears even before a teardown flow exists; it generates a cleanup flow paired to that seed. Toggle teardown off to preserve data for investigation.
• Set an API timeout — give a seed's setup/teardown API calls more or less time without affecting other steps.

Cofactor runs a seed automatically when a downstream step depends on one of its declared outputs, so you don't need extra steps just to make those values available.

05Best practices

• Keep seed flows small and focused on producing data, not asserting behavior — your standard tests own the assertions.
• Declare only the outputs your tests actually consume, with clear names.
• Pair every seed that creates persistent records with a teardown that removes them, so reruns stay clean.
• Pin a seed version when you need a test's setup to stay stable while the seed flow evolves.
• For disposable, fully isolated state, run against a sandbox instead of seeding shared environments.

06Related

• Flows
• Test suites
• Variables & macros
• Sandboxes