# Scheduled Tasks

The Scheduled Tasks view lets you create, manage, and monitor background jobs that the agent runs automatically. You can schedule a task to run once at a specific time, or set it to repeat on a cron schedule.

Tasks are useful for:

* Sending a daily summary report to a channel.
* Running a periodic check or data retrieval job.
* Triggering a one-off action at a specific datetime without manual intervention.

## How tasks work

When a task runs, the agent executes the task's **prompt** as if a user had sent it — the agent processes the prompt, optionally uses tools, and sends the result to the configured **output channel**.

### The task execution flow

Here's exactly what happens when your task's scheduled time arrives:

1. **Scheduler fires** — The scheduler sees it's time to run a task (e.g., 9 AM for a daily task)
2. **Create loopback message** — The scheduler creates a message from "System" containing your task prompt
3. **Message processing begins** — This message goes through the entire 11-step pipeline (see [Message Processing](https://neirth.gitbook.io/openlobster/architecture/processing-messages))
   * Validation (step 3) — Skipped for system-generated tasks
   * Context building (step 4) — Same as user messages: history, memory, permissions
   * Tool execution (step 9) — If your task needs to fetch data, query an API, etc., tools run here
4. **Response generated** — The agent produces an output
5. **Deliver to output channel** — Result goes to Telegram, Discord, a file, or wherever you specified
6. **Dashboard update** — You see the task result in Recent Logs

**Key insight:** Tasks are not a separate execution path. They go through the exact same pipeline as user messages, which means:

* Tasks have access to ALL tools (browser, filesystem, terminal, MCPs, memory)
* Tasks respect the same permissions system
* Tasks can use memory and reference past conversations
* Task results are saved to the database like any other message

**This is powerful:** A task can do everything a user could ask the agent to do, but on a schedule.

### Example: Daily summary task

You create a task:

* **Prompt:** "Generate a summary of yesterday's conversations about Q4 budget"
* **Schedule:** Daily at 9 AM
* **Output channel:** Telegram (your personal chat)

When 9 AM hits:

1. Scheduler fires
2. Creates message: "Generate a summary of yesterday's conversations about Q4 budget"
3. Message goes through the pipeline:
   * Loads relevant memory nodes about Q4
   * Searches recent conversation history
   * Uses memory search tools if needed (tool execution)
4. Agent generates: "Yesterday, 5 conversations mentioned Q4. Budget discussions included..."
5. Message sent to your Telegram
6. You see it in Recent Logs

If your task needed to query a database or fetch data from an external system, an MCP tool would handle that in step 3.

{% hint style="info" %}
The scheduler must be enabled for tasks to run. Verify that **Scheduler Enabled** is set to `true` in **Settings > Scheduler Configuration**.
{% endhint %}

## Pages in this section

* [Create, Edit & Delete Tasks](https://neirth.gitbook.io/openlobster/tasks/create-edit-delete)
* [Scheduling Formats](https://neirth.gitbook.io/openlobster/tasks/scheduling)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://neirth.gitbook.io/openlobster/tasks.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.