Flow Scheduling

While many Agent Flows are triggered manually during a chat session, RealTimeX also supports powerful Background Scheduling. This allows you to run flows automatically on a recurring basis without any manual intervention.

Scheduling Modes

You can configure the schedule for a flow in the Start Block settings.

1. Manual (Default)

The flow only runs when explicitly called by an agent or a user in chat.

2. Interval

Runs the flow repeatedly after a set amount of time.

• Options: Seconds, Minutes, Hours, Days, Weeks, Months.
• Example: "Run every 30 minutes" or "Run every 2 days at 9:00 AM".

3. Schedule (One-time)

Sets a flow to run at a specific date and time in the future.

• Example: "Run on 2026-12-25 at 00:00".

4. Custom Cron

For advanced users, you can use standard Cron expressions to define complex recurring schedules.

• Example: 0 0 * * 1-5 (Run at midnight every weekday).

Timezone Handling

RealTimeX uses UTC as the default timezone for scheduling. However, you can specify a custom timezone in the Start block configuration to ensure the flow runs at the correct local time for your region.

Execution History & Monitoring

Scheduled flows run in the background. You can monitor their performance and results through:

• Flow Logs: Check the "Logs" tab in the flow builder to see a history of every run, including successful completions and any errors.
• Calendar Integration: If your flow creates calendar events, you can see these in the RealTimeX workspace calendar.
• Notifications: You can build flows that send an email or Slack message (via MCP) upon completion or failure.

Best Practices

• Rate Limiting: When using short intervals (e.g., every 60 seconds), ensure your LLM and API providers can handle the frequency.
• Error Handling: Build your flows robustly. Use Conditional Blocks to check if an API call failed and stop the flow or send an alert.
• Overlap Protection: By default, RealTimeX prevents a new instance of a scheduled flow from starting if the previous one is still running.