Set a reminder, schedule a task that combines research, data and browser actions, or run a supported Replay again. The extension runs the task while Chrome is running and your device is awake.

Create a schedule

1. Open the toolbar’s Tools menu and choose Schedule → Create → Scheduled task.
2. The panel closes and puts /schedule in the chat input. Describe what should happen, when, and how to notify you, then send the request.
3. Answer any follow-up questions in the conversation.
4. Review the setup in the conversation and confirm when asked. Open the created task in Schedule → Edit to check its timing, materials, model, tools and notification channels.

You can also type /schedule in the AI Chat input, choose the command and add your request. You can create a schedule without switching to Agent S6. To change it later, go back to Schedule in the toolbar’s Tools menu.

For a reminder:

/schedule Remind me every weekday at 5 pm to update my reading notes. Use a browser notification.

For a combined task:

/schedule Every Monday at 9 am, read these course pages, compare the prices with the attached table, and add the changes to my Course tracker Google Sheet.

Attach the pages and table before sending. For a task that writes to Google Sheets, connect the intended account and turn on Apps. Write the request with the sites, conditions and destination the future run should use.

After creation, review Model and tools and Materials in the task editor. Those saved settings determine what the future task can use. Changes to another chat’s attachments or tool switches do not update this task.

Built-in schedule and monitor notifications start with the browser channel. Choose email, webhook or ntfy under Notify via in the task editor after setting up the channel in Notifications.

Choose the task type

Type	What runs
Reminder	Sends a notification, with any link-opening action you set. It starts no AI task.
Agent run	Runs the saved request with its selected materials and tools. It can combine reading, research, extraction, calculations and app actions, and use browser automation when enabled.
Replay skill	Runs a supported saved workflow with its configured inputs.
Page monitor	Checks a page and notifies you when its condition is met. See Page monitors.

To schedule a saved workflow, use Set as scheduled task from a supported Replay.

Set the materials, model and tools

In Edit task, expand Materials to add or remove source pages, uploaded files, referenced conversations, Knowledge Base material or Spaces.

Material	What a future run uses
Page selected for the task	The page at its saved URL, read when the task runs.
Uploaded file, image or referenced conversation	The version attached to the task when you saved it.
Space	The material available in that Space when the task runs, using its saved page-reading preference.
Knowledge Base documents or collections	The saved content available when the task runs.

Review Model and tools to choose the model and allow web search, browser automation, Apps, MCP, Skills and Commands. Choose a model to keep it fixed for this task, or select Follow current model settings to use the model selected in Settings when the task starts.

An app task retains the account selected when you configured it. If you disconnect that account or change an MCP connection’s tools, review the task’s settings before running it again. Keep required pages signed in and referenced materials available.

Set timing and the browser window

Choose Run once or Repeat. The editor supports a specific time, a delay, and recurring intervals or weekdays. After you save, check Next run to confirm the time.

For browser tasks, Where to run offers:

• Visible window: opens a separate visible window. Choose it for tasks that need to see the page.
• My current window: opens a tab in the window you are using, which may come to the front.
• Hidden window: runs out of sight. Steps that rely on screenshots or seeing the page may fail.

For a scheduled Replay skill, review Unattended action authorization and allow only the actions you want that workflow to take while you are away. An unchecked action can block the Replay.

An Agent run can wait for an answer, confirmation or help with the website. Open its run conversation to respond or continue.

A scheduled Agent run pauses after 10 minutes of active work. Select Continue in its conversation to allow another 10 minutes. A run that remains paused or waiting for you expires after 24 hours. Check the status before treating its work as complete.

If Chrome was closed or asleep

Late run recovery decides whether a delayed task runs once the browser is back. Within the window you set, Starizon runs one catch-up and skips older missed runs. It does not replay each occurrence Chrome missed.

When timing matters, check the last-run status. If a notification comes from a catch-up, the task ran at the catch-up time, not the scheduled one.

Manage and check runs

Open the task menu to Edit, Pause, Resume, Delete or open Run history. Use Run now to test a task without waiting for its next scheduled time.

Pause stops future runs. To stop a run already in progress, use Stop run. Stopping leaves in place any action already completed on the website.

Open Run history to inspect Agent and Replay runs, including ones that are running or waiting for you. Select an Agent run to open its conversation, inspect the tools and files, and answer a pending question. For page checks, open the monitor’s history. A task marked Needs attention has a page, account, model or instruction problem to fix before it can continue.

Plan limits and costs

Free supports two active scheduled tasks and two active monitors, counted separately. Paid plans support twenty active schedules and monitors in total.

On Free, the shortest recurring interval is one minute for reminders, fifteen minutes for non-AI tasks and thirty minutes for AI tasks. Paid plans allow shorter intervals, within the settings the editor supports.

Free includes 20 scheduled Agent tasks per day and 60 per month for tasks that use tools. After that, scheduled tasks can use any chat tasks you have left. A run combining several tools counts as one task. Browser steps and Replay have separate limits. If both task limits have been reached, Starizon can wait until a limit resets before running the task.

Model requests and Starizon services also use Starizon credits or are billed by your own provider. See Plans and credits for costs and task limits.