Skip to main content

Test automations

Try an automation before you rely on it. Use Manual run for a real run on records you choose, or Simulated run to check the steps without changing any data.

Written by Leo

You can test an automation in two ways. A manual run does the real work. A simulated run only walks through the steps. To start either, you need to be a workspace admin of the workspace the database app is in.

The automation editor of Expense approval with the status Paused below the name and Manual run, the runs button, and ••• in the header

Note:

  • A manual run has real effects. Records change, comments appear, and emails go out. It also counts toward your usage.

  • A manual run follows your filters. A manual run passes no field changes, so Has changed and Has changed to don't match and the record is skipped. To test such a filter, change the field on a test record so a real run starts, or use Is while you test.

  • A paused automation can still run manually. Pausing doesn't protect against Manual run.

  • A simulated run still calls other services. Requests from Send HTTP request, Send API request, and the http client in Execute script are sent for real. In a script, check current_workflow_is_simulation to skip such calls.

Start a manual run

  1. Open the automation in the editor.

  2. Select Manual run. Hover over it to see Start a manual run. The shortcut is cmd/ctrl + J. You also find Manual run in the ••• menu in the header of the editor.

  3. Next to Run on, choose Selected records, Selected view, Number of records, or All records. With Selected records, pick records from the list. Each shows the record name of its database app and the workspace, for example Expense in Accounting.

  4. Select Start run. If you have unsaved changes, the button reads Save and start run and saves them first.

You can't start a manual run on a draft record.

The Manual run menu with Run on selected records, a record search, Start run, and a list of expenses

Selecting Manual run opens Run on selected records with a record search, Start run, and the list of expenses; choosing Client dinner marks it, and the pointer rests on Start run

For triggers without a record, such as On a scheduled time or interval, the run needs no record. For When a webhook is received, the run uses data from a webhook. Next to Run on, choose Last payload to reuse the last data received, or Custom payload to paste your own.

To run on all records of a database app, read Run an automation on existing records →.

Start a simulated run

  1. Click ••• in the header of the editor.

  2. Select Simulated run. The shortcut is cmd/ctrl + K.

  3. Next to Simulate on, choose the records, the same way as for a manual run.

  4. Select Start simulation.

A simulated run skips the filter and doesn't change data. Use Manual run to test the filter. Actions that would change something, such as updating a record, adding a comment, or creating a record, only pretend to do it. Their steps still show Success, but nothing changes. The trigger and the filter show Skipped. Use it to check the structure of an automation and the values you insert. A simulated run doesn't send emails, but requests to other services go out for real.

The ••• menu in the editor header with Save, Undo, Redo, Full width, Manual run, Simulated run, Cancel runs, Export runs, Delete, and the line Last modified by

Check the result

  1. Click ▷ next to Manual run. Hover over it to see View all runs. You can also press cmd/ctrl + 1.

  2. Find your run. Manual runs are listed as Manual run on with a date and time. Runs started by the trigger are listed as Run on. Simulated runs are listed as Simulated run on with a date and time.

  3. Select the run to see each step and its status, such as Success or Skipped.

Learn more in Automation runs →.

Clicking the runs button next to Manual run to open the Runs panel, opening a failed run, and expanding the failed step Create a record to read its error

Test an automation safely

  • Use a few records that don't matter.

  • Start with Simulated run, then use Manual run on one record.

  • Set the automation to Paused while you test, so real changes don't start it. Learn more in Active, paused, and broken automations →.

Learn more

FAQs

Can I run a paused automation manually?

Yes, Manual run works while the automation is Paused. Only the trigger is paused.

Can I restart a failed run?

No, a failed run can't be restarted from the run itself. Fix the cause, then start a Manual run for the affected records.

Does a simulated run send emails?

No, a simulated run doesn't send emails, and it doesn't change records. Requests to other services, for example with Send HTTP request, are still sent.

Are you a developer using the Tape API for automations?

The automation API has its own documentation for developers. Full reference: Run and simulate in the developer docs →

Still have more questions? Find answers and get help from community experts in the Tape community →.

Did this answer your question?