Describe the outcome
Name the inputs, destination, trigger, time zone and exceptions. Say what the workflow must not do. Describe it here and Whenever writes it, start from a template, or bring the brief to your coding agent.
Describe the process in plain words and Whenever’s AI writes it as a typed workflow. Test it against real inputs, publish the trigger, and from then on it runs as code. The only model call in a run is one you chose to put in a step.
Every weekday at 09:00 Europe/Vilnius, read our customer-onboarding sheet. Select rows marked blocked, grouped by owner. Post one summary to the onboarding Slack channel. If there are no blocked rows, post nothing. If an owner is missing, flag that row for review. Do not edit the sheet.
Example brief, not executable SDK code. Connect the actual sheet and channel, review permissions, and verify the generated workflow before publishing.
Explicit Europe/Vilnius time zone.
Use the authorized spreadsheet connection.
Blocked only. Missing owners flagged. Empty means stop.
One destination. No spreadsheet writes.
The brief is the beginning, not the acceptance test. Keep the important decisions visible at each handoff.
Name the inputs, destination, trigger, time zone and exceptions. Say what the workflow must not do. Describe it here and Whenever writes it, start from a template, or bring the brief to your coding agent.
Whenever workflows are TypeScript modules with a trigger manifest and a runnable workflow. Inspect branches, named steps and integration calls instead of relying on the prompt alone.
Run manually against controlled data and a test destination. Try the ordinary case, the empty case and the missing-field case. Inspect the actual output before connecting live actions.
Choose a schedule, webhook or supported provider event. Authorize the required connections, check run history and keep an owner responsible for exceptions.
Use this as the review checklist for your first workflow.
The same plans and allowances as our pricing page. Model usage and external services belong in your estimate too.
$0/mo
100 runs per month
One active workflow, no credit card.
* One run = one workflow execution, start to finish. Free AI building and editing have weekly usage limits.
It helps turn a plain-language description into an automation. In Whenever, the result is a typed workflow with explicit triggers and logic. You still review its actions, authorize connections and test real inputs before publishing.
You can start with a brief, a template or your coding agent. The workflow itself is TypeScript, so code remains available to inspect and change. Someone should review the logic and permissions before it handles live business data.
No. AI can help write a workflow that uses only explicit rules and API calls at runtime. If the process needs interpretation, add a bounded AI step and validate its output before downstream actions.
The documented trigger builders include manual runs, schedules, one-time runs, webhooks and managed provider events. Provider events and integration actions depend on the connections available and authorized in your workspace.
You can inspect and adapt the TypeScript logic. Do not assume it is a drop-in deployment elsewhere: integration ports, triggers, credentials and step execution depend on the runtime.
Sources reviewed September 24, 2026.
Make the rules explicit. Test the edge cases. Publish when the result is ready.