Scheduled Tasks: Cron-Style Automation
Configure your OpenClaw agent to run tasks on a recurring schedule, from hourly health checks to daily report generation.
What You Will Get
By the end of this guide, your OpenClaw agent will execute tasks automatically on a fixed schedule without any manual trigger. You will configure cron expressions to define when tasks run and what the agent should do each time.
Scheduled tasks are perfect for recurring work like generating daily summaries, checking external service health, cleaning up stale data, or sending periodic updates to a channel. Instead of relying on a user to initiate these actions, the agent handles them autonomously at the times you specify.
You will also learn how to handle missed schedules, manage overlapping executions, and review task history. The result is a reliable automation layer that keeps your workflows running on time, every time.
Step-by-Step Setup
Follow these steps to set up your first scheduled task.
Open the Scheduling Panel
Navigate to your agent's settings in the RunTheAgent dashboard and select the Scheduled Tasks tab. This panel shows all existing schedules and lets you create new ones. If you have no schedules yet, you will see an empty list with a button to add your first task.
Create a New Scheduled Task
Click the Add Schedule button and give your task a descriptive name like 'Daily Summary Report' or 'Hourly Health Check'. A clear name makes it easy to identify the task in logs and alerts later.
Set the Cron Expression
Enter a cron expression to define the schedule. For example, '0 9 * * *' runs the task every day at 9:00 AM UTC, and '*/30 * * * *' runs it every 30 minutes. The panel includes a preview that shows the next five execution times so you can verify your expression is correct.
Define the Task Prompt
Write the prompt that the agent will execute each time the schedule fires. Be specific about what the agent should do and what output to produce. For example, 'Generate a summary of all conversations from the past 24 hours and send it to the reports channel.' The agent treats this prompt as if a user sent it.
Configure Output Destination
Choose where the agent should deliver its results. Options include sending the output to a connected channel, storing it in the agent's memory, or posting it to a webhook endpoint. You can select multiple destinations if you need the output in more than one place.
Set Overlap and Retry Policies
Decide what happens if a previous execution is still running when the next schedule fires. You can skip the new execution, queue it, or allow them to run in parallel. Also set a retry count for failed executions so transient errors do not cause missed reports.
Activate and Monitor
Toggle the schedule to active and wait for the first execution. Check the task history log to confirm it ran at the expected time and produced the correct output. Adjust the cron expression or prompt as needed based on the results.
Tips and Best Practices
Use UTC for All Schedules
Cron expressions run in UTC by default. Convert your local time to UTC when setting schedules to avoid confusion, especially if your team spans multiple time zones.
Start with Low Frequency
Begin with daily or hourly schedules and increase frequency only when needed. Running tasks too often wastes tokens and can hit rate limits without adding value.
Include Error Handling in Prompts
Instruct the agent on what to do if it encounters an error during a scheduled run. For example, 'If the data source is unavailable, log the error and retry in 10 minutes.' This prevents silent failures.
Review Task History Weekly
Check the execution history at least once a week to spot patterns like increasing run times, frequent retries, or unexpected outputs. Early detection of issues prevents larger problems down the line.
Frequently Asked Questions
Related Pages
Ready to get started?
Deploy your own OpenClaw instance in under 60 seconds. No VPS, no Docker, no SSH. Just your personal AI assistant, ready to work.
Starting at $24.50/mo. Everything included. 3-day money-back guarantee.
