Skip to content
Docs
flero.ai

Testing a node

You can run a single node in isolation while you build, instead of executing the whole workflow. This shortens the develop-test-fix loop dramatically, especially for nodes that talk to slow or rate-limited APIs.

Two ways to test

1. "Test" button in the inspector

Every action node has a Test button at the bottom of its inspector. Clicking it:

  1. Resolves all expressions in this node's config against the last execution's data (or, if there's no last execution, against empty data and any pinned inputs).
  2. Executes only this node.
  3. Renders the result in the Output section of the inspector.

📸 Screenshot needed: editor__test-node-output.png, HTTP Request node inspector with the Test button highlighted and the Output section below showing a JSON result.

If the test fails, the inspector shows the error inline with the same red styling you'd see on a full-workflow run.

2. Run-from-here

Right-click any node → Run from here. Flero executes that node and every node downstream of it, treating upstream nodes as if their last-execution outputs were the current inputs.

Useful when you've just changed a node mid-chain and only want to verify everything downstream.

Where the input data comes from

A test run needs some data on every input port. Flero resolves this in priority order:

  1. Pinned data on the upstream port (see Pinning data below), if set, this wins.
  2. The most recent execution's output for the upstream node, if there's been a run.
  3. An empty object otherwise.

This is why your first test on a brand-new workflow often produces an empty-looking result. Run the whole workflow once (or pin sample data) and subsequent tests will resolve everything.

Pinning data

Pinning lets you hard-code an input for development without changing the upstream nodes.

  1. Run the workflow once (or paste sample JSON manually, see step 3).
  2. In the upstream node's Output preview, click the Pin icon (📌).
  3. Pinned data is now used as the input for every subsequent test until you unpin.

You can also paste sample JSON directly:

  1. Open the upstream node's inspector → Output → Pin custom data.
  2. Paste a JSON object or array.
  3. Save.

Pinned data is not persisted to executions, it only affects test runs in the editor.

📸 Screenshot needed: editor__pin-data.png, Output panel of an HTTP node with the Pin icon highlighted and a "Pinned" badge visible.

Sandbox mode for tests

If your workflow talks to real external services (Slack, Salesforce, billing APIs), use sandbox mode to swap them for mock responses while testing:

  1. In the bottom execution chrome, toggle Sandbox on.
  2. Test runs from this point use the Sandbox mock pipeline.

This is invaluable when iterating on the shape of your data, no risk of accidentally sending an email to a real customer.

Tips & gotchas

  • Test is per-node; it doesn't follow If or Switch branches downstream. To test branching, use Run from here instead.
  • Side effects are real unless sandbox mode is on. A test of a Slack node will post to your real Slack workspace. Be deliberate.
  • Tests don't update the execution history. They're considered editor-local development runs. If you want a row in /logs, do a full workflow run.
  • Pinned data persists across save, useful for fixtures. Forget to unpin before going live and your workflow could behave unexpectedly. The bottom chrome shows a Pinned data active indicator when any port has pins.
  • Long-running tests can be cancelled by pressing the Stop button that appears in place of Test.

Found something out of date? This page lives in the Flero docs content set.