Skip to main content

Triggers

A trigger makes an agent run without a human typing in chat. Find them under Integrations → Triggers.

What you can trigger on

Trigger typeWhen it fires
ScheduleOn a cron expression — every hour, every Monday at 9am, etc.
WebhookWhen an external system POSTs to a unique URL
EventWhen something happens in the platform (new task, agent-team run completes, etc.)

The Triggers page has three tabs: Schedules, Webhooks, and Run History.

Triggers page with Schedules tab open showing scheduled agent runs and their last-run status

Setting up a trigger

Go to Integrations → Triggers → Add Schedule (or Add Webhook).

You'll pick:

  • Name — for your reference
  • Type — schedule, webhook, or event
  • Target — agent or agent team to run
  • Input template — the message or payload sent to the agent

Save. The trigger is now active.

Schedule triggers

Use a cron expression or pick a preset (every hour, daily at 9am, every Monday morning). Useful for:

  • Daily summaries
  • Weekly digests
  • Hourly health checks
  • Periodic data refreshes

The agent runs as a background task. Output goes to the agent's interaction log and to any output destination you've configured (email, Slack, webhook).

Create Schedule dialog: target picker (Agent / App), agent selector, name, description

Webhook triggers

The trigger gives you a unique URL. Anything that can POST JSON can call it — Zapier, n8n, GitHub Actions, your own backend.

The POST body becomes the input to the agent. You can shape it with the input template: "User {{user.name}} just signed up at {{user.company}}. Send them a welcome message.".

The webhook URL includes a secret token (HMAC-signed). Treat it like a password.

Webhooks tab: configured webhook with generated URL, HMAC-signed badge, agent target

Event triggers

Run an agent in response to something inside Agent Studio:

  • A task lands in Async Tasks — auto-pick it up
  • An agent-team run finishes — post the result somewhere
  • Another agent finishes — chain the next one

These let you build pipelines without external orchestration.

Test before enabling

Every trigger has a Test button. It runs the trigger once with a test input and shows you the agent's response. Use it before turning a trigger loose on real data.

Disabling vs deleting

A trigger you toggle off stops firing but stays in the list. You can re-enable it any time.

A trigger you delete is gone (well — soft-deleted, see Trash).

Trigger history

Open any trigger and click History to see:

  • Every fire, with timestamp
  • Whether the agent succeeded or failed
  • The input that came in
  • A link to the agent interaction log

If a trigger seems quiet when it shouldn't be, history is the first place to check.