Skip to main content
Choose Automation inside a colony. Its four sections handle different parts of running and supervising work.
Screenshots show the desktop interface with fictional example data. Names, messages, counts, and connection states are illustrative, not account defaults or evidence of live activity.
Colony Automation

​
Scheduler: arrange future runs

Use New to create a schedule for the colony. Review the schedule’s next fire time and state after saving. Force Run invokes the trigger now; Stop trigger disables future invocations, and Start trigger re-enables it. Read Scheduling for daily, weekly, interval, and advanced cron behavior, time zones, and local-machine requirements. Stopping a current worker does not disable a recurring trigger.

​
Notifications: hear when the colony needs you

Choose Set up, or Manage for an existing configuration. Notifications can use a supported Telegram or Slack destination.
  1. Connect or select the channel and destination.
  2. Review Allowed repliers. These are the user IDs allowed to drive the colony through replies; do not add people casually.
  3. Enable alerts for this colony.
  4. Set Alert me if stuck for to the available delay you want.
  5. Finish the setup and check that the section shows On with the intended destination.
The setup can guide you through connecting a Telegram bot or Slack app. Agent-assisted setup may involve browser actions, credentials, and a test message; review what you are authorizing before starting it. A connected messaging account alone is not the same as alerts being enabled for this colony. If alerts do not arrive, check the destination, enabled state, and channel setup errors.
Notification settings with the Alert me if stuck for control set to 15 minutes.

Choose how long the colony can be stuck before it alerts you.

​
Cloud: choose where the work runs

Use Cloud to deploy the colony to a cloud computer or open an existing deployment. Read the confirmation before changing where it runs. Cloud deployment and scheduling are separate decisions. A schedule does not automatically deploy a local colony, and local execution depends on your app and computer remaining available. See Local and cloud for deployment and moving back to local execution.
Review the cloud confirmation before starting cloud execution.

Review the cloud confirmation before starting cloud execution.

​
Workers: inspect delegated execution

Workers carry out tasks delegated by the queen. The active list shows ongoing work; History shows previous workers and their outcomes. Expand the history or show more entries when needed. Select a worker to inspect its goal, task, timing, results, and available transcript. Use All workers to return. Skills & tools opens the worker capabilities view. If a result is unexpected, compare the worker’s instructions with its tool activity and reported result. A failed worker does not necessarily mean the entire colony failed; the queen may have other work or a recovery path.
Inspect active workers and expand History to review earlier execution.

Inspect active workers and expand History to review earlier execution.

​
Stop the right thing

Stop this worker interrupts one worker. Stop all interrupts the active workers listed in this section. Neither action undoes external changes already completed, and neither disables future scheduled triggers. Use the colony’s running control to pause colony execution, and Stop trigger for a schedule you no longer want. These are different controls with different scopes.