Recurring tasks
Recurring tasks fire on a schedule and keep firing until you stop them or the end date is reached.
![]()
Create a recurring task
- Start a new task — see One-time tasks for the basics (title, agents, prompt).
- Toggle the Recurring switch.
- Set the recurrence:
- Frequency — Daily / Weekly / Monthly.
- Interval — every N days/weeks/months.
- Start date — when recurrence begins.
- End date (optional) — when recurrence stops.
- Click Create Task.
Common patterns
| Pattern | Frequency | Interval |
|---|---|---|
| Daily report | Daily | 1 |
| Weekly summary | Weekly | 1 |
| Monthly analysis | Monthly | 1 |
| Bi-weekly update | Weekly | 2 |
What happens at run time
- First execution runs at the configured start date + time.
- Subsequent executions follow the recurrence pattern.
- Each run creates a separate execution history entry. See History & logs.
- The task stays active until you disable it or the end date is reached.
Time zone & date/time picking
![]()
- All times are in your workspace timezone.
- Verify timezone in Settings → Workspace settings.
- For distributed teams, consider the receiving team’s working hours, not yours.
Tips
- Schedule resource-heavy tasks off-peak (early morning, weekends).
- Group related recurring tasks so they run sequentially; this is easier to monitor than scattered jobs.
- Monitor the first few executions — recurring tasks fail silently if you don’t watch them.
- Use an end date for time-bounded campaigns. Otherwise the task runs forever until someone notices.
- Avoid prompts that reference “today” without a date — model interpretations of “today” can be inconsistent across runs.
Related
- Priorities & status — how priority interacts with scheduled execution.
- Execution & monitoring — what to watch.
- External triggers — when “every Tuesday at 9am” isn’t the right trigger.