> ## Documentation Index
> Fetch the complete documentation index at: https://docs.rundesert.com/llms.txt
> Use this file to discover all available pages before exploring further.

# What are triggers

> The four ways an automation starts.

The automation is the work. The trigger is what starts it.

Every automation has at least one trigger. You never wire one up yourself. You
tell the agent when the job should run, and the agent sets it up.

## The four kinds

| Trigger                                         | It runs when                    | Reach for it when                                        |
| ----------------------------------------------- | ------------------------------- | -------------------------------------------------------- |
| [**Scheduled task**](/triggers/scheduled-tasks) | The time you picked comes round | You want a daily digest, a nightly sync, a weekly report |
| [**Webhook**](/triggers/webhooks)               | Another app calls its URL       | GitHub, Stripe, or Zapier should kick off the job        |
| [**Incoming email**](/triggers/incoming-email)  | A message lands in your inbox   | The work arrives as mail                                 |
| [**Manual**](/triggers/manual)                  | You press a button              | You want it now, or you want to test it                  |

## One automation can use several

Triggers stack. You do not pick one and give up the rest.

The same automation can run every night at nine, run again when a webhook fires,
and run once more when you press the button. All three paths hit the same code.
The trigger only decides when the work starts, never what the work does.

This helps more than it sounds like it would. You can always press the button to
watch your automation run while you build it. And when you add a schedule three
weeks later, you do not rebuild anything.

## Your workspace sleeps, and that is fine

Your workspace pauses when nothing is happening. That is why you can leave a
dozen automations sitting around without paying for idle time.

It also means a timer running inside the machine would die with it. So the
platform keeps the clock for you and owns the public URLs. When your schedule
ticks, when someone calls your webhook, or when mail arrives, the platform wakes
your workspace and hands over the event. You lose nothing while it sleeps.

<Note>
  Waking takes a few seconds. A trigger can land 10 to 60 seconds after the event
  that caused it. That is normal. Do not build anything that needs to fire at an
  exact second.
</Note>

## Seeing what you have set up

Open the **Automations** tab. Every trigger shows up there on the automation it
belongs to, with its type, a plain-English label, and the URL if it uses a
webhook.

Check that tab whenever you wonder what your workspace will do on its own.