Open App

Scheduled Assigning

Schedule local agent launches or choose run modes for connected task agents

After assigning an agent to a task, choose how runs are dispatched.

Scheduled local launches

In the desktop app, open a task’s Open in AI coder menu and choose Schedule launch…. Select the agent, working folder and terminal where applicable, then choose:

  • Date/time: launch once at the selected date and time, in this laptop’s timezone.
  • Next routine start (once): launch once at the task’s next linked routine start, respecting date-specific overrides.

Runs on this laptop only. Keep MONOid open and your laptop awake at the scheduled time. Minimized is okay. Missed runs won’t start automatically.

Stay signed in and online: MONOid loads current task details and your current prompt template at launch time. This opens your usual local agent session; login or permission prompts can still need your input. Completed, canceled, deleted, or inaccessible tasks are skipped. Routine timing or linkage changes require rescheduling.

Task controls show Edit, Cancel, and Run now. A successful manual launch replaces the pending local launch for that task. Settings > Local agents > Scheduled launches lists schedules and recent outcomes on this laptop. After reopening, one grouped notification links to missed or failed launches for review and rescheduling.

Launch sent confirms that the command was handed to the tool, not that the task is complete. If MONOid stops during dispatch, Launch status unknown means the agent may already be running. Check the agent before choosing Launch again. Schedules are local to your account and this laptop; they do not sync to another computer or execute from the browser.

Run modes

  • Run now (immediate): launch an agent run immediately.
  • Run on routine block (routine_trigger): schedule run mode so execution is tied to the task's routine block timing.

How to set run mode

On the task agent controls:

  1. Open the run menu.
  2. Choose Set Run now or Run on routine block.
  3. For scheduled mode, keep the task linked to a routine block so the timing is meaningful.

What you can monitor

Task agent cards show:

  • Run status (queued, launching, running, finished, error, expired, canceled)
  • Run link / PR link
  • Branch name
  • Summary or error details
  • Needs review badge

Provider note

Run controls are currently available for providers with integrated run flows in task controls (Cursor, Codex, OpenClaw).