Workflow Test Flow
How to test a workflow end-to-end with Jarvis-driven test runs before enabling it on live triggers.
Before you flip a workflow live, the Test Flow tooling lets you fire it manually against a sample payload and watch every step execute against real services. This is the safest way to confirm a flow does what you expect before it starts spending credits or sending live messages.
When to use a test run
- After building a new workflow in the canvas.
- After editing a live workflow's logic or destination.
- When debugging a workflow that previously succeeded but stopped producing the right output.
Running a test
- Open the workflow and click Test in the editor toolbar.
- Pick a sample trigger payload — either a recent real event (from prior runs) or a synthetic one Jarvis can generate for you.
- Run the test. Each step executes in order against real integrations, so destinations like Slack, Sheets, or webhooks will receive the test output unless you explicitly route to a sandbox channel.
- Inspect the result in the Output tab — it appears alongside live runs but is flagged as a test execution.
What a test run does and doesn't do
- Does — execute every step against live integrations, consume credits where credit-bearing actions are involved, and write to the Output tab.
- Doesn't — bypass auth, mock integration responses, or skip rate limits. If a downstream API is rate-limited, the test will hit the same limit.
Best practices
- Test with a representative payload, not the smallest possible one. Edge cases (empty arrays, long text, missing fields) are the runs that fail in production.
- Route destinations to a test Slack channel or sandbox sheet for the first few runs.
- Once a test run looks correct, publish the workflow and enable the live trigger.