Proactive Agents: Automating Your Life with OpenClaw Cron Jobs

2026-02-08•Mintu

In our last post, we talked about how OpenClaw solves the "amnesia" problem using persistent memory. But there's another major limitation with standard chatbots: passivity.

Standard AI models are reactive. They sit there, frozen in time, waiting for you to type "Hello." If you forget to ask for your morning briefing, you don't get it. If you forget a meeting, your AI won't remind you—because it effectively doesn't exist until you hit Enter.

OpenClaw changes this with the Cron system.

By integrating a scheduler directly into the Gateway, OpenClaw empowers your agent to wake up, look around, and take action without you saying a word.

The Shift from Tool to Partner

A tool waits for a hand to guide it. A partner takes initiative.

With OpenClaw's Cron feature, your agent can:

  • Wake up at 7:00 AM to scan your emails and prepare a summary.
  • Nudge you in 20 minutes to check the oven.
  • Monitor a website every hour and alert you only if a specific keyword appears.

This transforms the relationship. You're no longer just prompting a model; you're orchestrating a lifecycle.

How It Works

The Cron system lives in the OpenClaw Gateway, the always-on service that manages your agent. It supports two powerful modes of operation:

1. System Events (The "Nudge")

This mode injects a message directly into your main chat session. It's perfect for reminders or context updates.

  • Scenario: You're deep in a coding session and tell your agent, "Remind me to push these changes in an hour."
  • Action: The agent creates a one-shot cron job.
  • Result: One hour later, the agent "wakes up" and types into your chat: "System Event: Reminder to push changes." It's just like a text from a friend.

2. Isolated Sessions (The "Background Worker")

Sometimes you want your agent to do heavy lifting without clogging up your main chat history. Isolated sessions spin up a temporary, fresh instance of your agent to perform a specific task.

  • Scenario: Every morning at 8 AM, you want a briefing on tech news.
  • Action: You schedule a recurring cron job with a specific prompt: "Search TechCrunch and Hacker News for top stories and summarize them."
  • Result: The agent runs this task in the background. If—and only if—it finds something worth reading, it delivers the summary to your preferred channel (Slack, Telegram, etc.).

Scheduling in Practice

You don't need to write complex code to set this up. You can use the openclaw CLI or simply ask your agent to do it (if it has the cron tool enabled).

Example: The Morning Briefing

openclaw cron add \
  --name "Morning Brief" \
  --cron "0 8 * * *" \
  --tz "America/New_York" \
  --session isolated \
  --message "Check my calendar and unread emails. Summarize my day." \
  --deliver

Example: The Quick Reminder

openclaw cron add \
  --name "Laundry" \
  --at "45m" \
  --session main \
  --system-event "Reminder: Move laundry to dryer." \
  --wake now

Why "Wake Mode" Matters

OpenClaw isn't just running a script; it's invoking an LLM. This costs money and compute.

The Cron system is designed to be efficient. You can choose:

  • Wake Now: Run the model immediately when the job fires.
  • Next Heartbeat: Queue the event and wait for the next natural interaction or scheduled check-in. This saves tokens by batching non-urgent updates.

Building a Proactive Life

The magic happens when you combine Memory and Cron.

Imagine an agent that remembers you're trying to drink more water (Memory) and schedules random check-ins throughout the day to ask if you've hydrated (Cron). Or an agent that knows you're tracking a specific stock and checks the price every close, logging it to a file.

This is the promise of OpenClaw: AI that lives with you, managing the background noise so you can focus on the signal.

Ready to automate your day? Download OpenClaw and start scheduling your first agentic workflow.

Back to Blog