> ## Documentation Index
> Fetch the complete documentation index at: https://docs.every.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Scheduled Tasks

> Schedule agent work to run as a recurring routine or a one-time reminder — built-in and custom tasks, managed from Agent or right in chat

## Overview

Schedule agent work to run as a recurring routine or a one-time reminder. Instead of asking Every something each time, turn on the tasks you want it to keep doing — a morning brief, a sweep of overdue invoices, or any custom request — and Every runs them for you, on your schedule.

> **[Open Scheduled Tasks](https://app.every.ai/agent/tasks)**
>
> Turn tasks on or off, set their cadence, and review past runs

**Highlights:**

* Every run uses full agent context and posts its results to its own thread, with undo history
* Recurring tasks run on a cadence you choose — daily, weekly, monthly, or a custom interval — with timezone-aware start times
* One-time tasks run once, at a specific date and time
* Results show up in the Activity Timeline, plus an optional email summary
* One agent task runs at a time per workspace, so scheduled work never overlaps
* A manual **Run Now** lets you test a task without touching its future schedule

> [!NOTE]
>
> Scheduled Tasks was previously called "Heartbeat" / "Automations." The old `/agent/heartbeat` and `/agent/automations` links now redirect to **Agent → Scheduled Tasks** at [https://app.every.ai/agent/tasks](https://app.every.ai/agent/tasks).

## Recurring

Recurring tasks run on a repeating cadence you set. A built-in Daily Brief is always available, and you can add as many custom recurring tasks as you like.

### Built-in tasks

| Task | What it does |
| - | - |
| **Daily Brief** | Your "chief of staff" morning readout of what moved overnight and what needs your attention — see [Daily Brief](https://docs.every.ai/scheduled-tasks/daily-brief.md). |

The built-in remains available in the task list. Turn each on or off and adjust its schedule and delivery. Existing Accounts Receivable Sweep tasks appear among your custom tasks; you can also create a custom task to review overdue invoices and draft reminders.

Automatic Network & Deals maintenance runs behind the scenes and is not a task you manage here. Manage Google discovery from your Google connection settings.

Prospecting is not a scheduled task. It runs automatically from the [Prospects](https://app.every.ai/prospects) page — see [Prospecting Agent](https://docs.every.ai/prospecting/prospecting-agent.md).

### Custom recurring tasks

Write your own recurring request and Every runs it on the schedule you set — for example, "Every Monday, summarize last week's revenue and open proposals." Use **Add Task**, describe the work, and pick a cadence.

## One-time

A one-time task runs once, at a specific date and time, then it's done — useful for a single reminder or a request you only need run once, like "Next Friday at 9am, check whether the ACME invoice was paid." Create one from **Add Task** the same way as a custom recurring task, but choose a specific date and time instead of a repeating cadence.

## Creating a task

Use **Add Task** from [Scheduled Tasks](https://app.every.ai/agent/tasks) (or ask Every Agent in chat) to create a custom recurring or one-time task:

1. Describe the request in plain language.
2. **@-tag** specific people, companies, or deals to scope the task to them, or **@-tag an active [skill](https://docs.every.ai/ai/memory-and-skills.md)** so the task follows that skill's guidance.
3. Choose **Recurring** (pick a cadence) or **One-time** (pick a date and time).
4. Save — Every runs it from here on.

> [!TIP]
>
> **Ask Every Agent:** You can also set up a task conversationally — "Every weekday at 8am, check my inbox for unpaid invoices and draft reminders."

## Cadence & scheduling

Every recurring task stores a local run time, so it fires at the right moment for your timezone.

| Cadence | Behavior |
| - | - |
| **Daily** | Runs every day (or on chosen days of the week) at a set time. |
| **Weekly** | Runs on a chosen day of the week at a set time. |
| **Monthly** | Runs on a chosen day of the month at a set time. |
| **Custom interval** | Runs every *N* hours or days. |
| **One-time** | Runs once at a specific date and time. |

> [!NOTE]
>
> Tasks that analyze history (like the Daily Brief) keep their lookback window in sync with cadence — a daily task reviews the last day, a weekly task the last week — so coverage doesn't overlap or leave gaps.

## Managing tasks from chat

Every agent thread has a **clock icon** that opens a popup listing the tasks connected to that conversation, so you can check status, adjust cadence, or turn a task off without leaving the chat. Active [skills](https://docs.every.ai/ai/memory-and-skills.md) can also be @-tagged directly in chat, the same way you'd tag them when creating a task.

## Results, threads & undo

* **Its own thread.** Each run posts a summary, links, and any drafts into a dedicated agent thread. Open it from [Every Agent](https://app.every.ai/agent) to read the full result or follow up directly.
* **Undo history.** Runs that change records capture an undo snapshot, so you can revert a scheduled change right from the thread.
* **Activity Timeline.** Upcoming runs preview on the [**Activity**](https://app.every.ai/activity) page; completed runs show success or failure badges with quick links back to the originating thread.
* **Optional email.** Turn on the email summary for any task to also receive a recap by email after each run.

> [!TIP]
>
> Use **Run Now** on any task to validate a fresh setup on demand. A manual run uses the same executor as the schedule but leaves your future runs untouched.

## Good to know

> [!WARNING]
>
> **One task at a time per workspace.** If you trigger a run while another agent task is already in flight, you'll see "Another agent task is running for this workspace" — wait for the current run to finish (check the Activity Timeline or the latest agent thread) before retrying.

* **Tasks use AI credits.** Each run consumes credits based on the work the agent does. If your credits are exhausted when a run is due, it's skipped — see [Credit Usage & Errors](https://docs.every.ai/plans/credit-usage-and-errors.md) for exactly what happens (including why a skipped run can show as "Completed").
* **Team scope.** In a multi-member workspace, your Daily Brief is private to you, while custom task definitions are shared with the team. Cards show a **Private** or **Team** badge so you know who can see each one.

## Next steps

> **[Daily Brief](https://docs.every.ai/scheduled-tasks/daily-brief.md)**
>
> Configure your chief-of-staff morning digest

> **[Prospecting Agent](https://docs.every.ai/prospecting/prospecting-agent.md)**
>
> Runs automatically from the Prospects page, with no schedule to manage

> **[Sales Pipeline](https://docs.every.ai/prospecting/pipeline.md)**
>
> See how reviewed deals move through your pipeline

> **[Credit Usage & Errors](https://docs.every.ai/plans/credit-usage-and-errors.md)**
>
> What tasks cost and what happens on errors or exhaustion

> [!NOTE]
>
> **Need help?** Contact support at [hello@every.ai](mailto:hello@every.ai)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.