Workflows

A workflow is a job the assistant sets up once and then runs again on its own, on a schedule or whenever something happens. You do not need to open the chat and ask each time.

​

When to use a workflow instead of a one-off job

Ask for a one-off job when you want an answer right now and will not need it again. Ask for a workflow when you want the same job to repeat, or to fire automatically when something happens.

Check this product’s price every morning at 9am and tell me if it drops below $40.

Whenever a new message arrives on my rented number, forward the code to my webhook.

If you are not sure yet whether something should recur, say so. The assistant can build it as a one-off first and turn it into a workflow later.

​

How the assistant builds a workflow

Building a reliable workflow takes more than writing a prompt. The assistant walks through four phases in the same chat, and tells you which phase it is in. Along the way the assistant asks you a small set of questions up front, like cadence, which device, and how results should be checked, so it does not have to interrupt you again later.

​

Triggers

A workflow needs exactly one trigger, chosen when you set it up. For a time-based schedule, the assistant asks which timezone to use. Manual triggers do not need one.

​

What a run produces

A run can produce a downloadable file, a short summary posted back into the chat, or both, depending on what you asked for. For result checking, you choose how thorough the assistant should be: it can trust the automation’s own checks, review the result itself, or do both, flagging only what looks off. Ask for one of these when you set the workflow up, or leave it to the assistant’s judgment for a simple job. A workflow can also send its result straight to a webhook endpoint you own, as one of its steps. That is different from the Notify on success / failure switch in the workflow’s dashboard settings, which sends a pass/fail notification on every run. The assistant cannot turn that switch on for you; set it in the dashboard. See Notifications and Webhooks for both mechanisms.

​

Running the same job across many phones

Ask the assistant to run a workflow, or a one-off job, across several devices at once.

Run this same check on all five of my cloud phones.

The assistant confirms it actually has that many devices ready before starting, runs up to eight of them in parallel, and retries any device that failed once before reporting back. You get a result per device, not just “mostly done.”

​

When a run fails or repairs itself

Every workflow watches its own health by default. If a run fails for a fixable reason, the assistant can automatically diagnose and repair it within a limited number of attempts, then try again. You can ask it to change how many attempts it gets, or turn automatic repair off for a given workflow. A failure that is not fixable this way, such as a login the assistant cannot get past, is not silently retried forever. The assistant reports it and asks what you want to do next.

​

Inspecting past runs

Every run is recorded. Ask the assistant to show you the history of a workflow, what happened in a specific run, or why the last one failed, and it reads that back to you.

​

Stopping or deleting a workflow

Ask the assistant to delete a workflow it built, and it removes everything it created for that automation: the trigger, the steps, and the flow. Pausing or disabling a live workflow without deleting it is done in the workflow’s settings in the dashboard.

​

Cost

Building a workflow uses credits the same way a normal job does, since the Learning, Testing, and Deploy phases each involve real work. Every time a deployed workflow fires afterward, that run also draws from your credit balance. See Limits and costs for how usage is measured.

​

Related