# Manager (/docs/verdent-manager/core-features/manager)

> Built for everyone — tell Manager what you want, and watch it deliver



Built for everyone, not just engineers. Manage products, workflows, and AI agents in one place with almost zero setup. You don't need a master plan to start or learn how to use AI coding. Just tell Verdent what's on your mind, and watch Verdent magically turn them into a working app.

***

## You don't have to manage the work [#you-dont-have-to-manage-the-work]

You don't need to sit in front of the computer, break down tasks, set priorities, or babysit progress. Hand the goal to Manager — it understands what you want, drives execution end to end, and delivers the result directly back to you.

<Steps>
  <Step title="Tell Manager the outcome">
    Describe what you want in plain language. No structure, no ticket template, no prompt engineering.
  </Step>

  <Step title="Manager takes over">
    It plans the work, breaks it down, dispatches agents, and runs everything in parallel.
  </Step>

  <Step title="You get the result">
    Come back to a finished deliverable, send feedback, or ask for the next thing.
  </Step>
</Steps>

***

## What Manager can do for you [#what-manager-can-do-for-you]

<CardGroup cols="2">
  <Card title="Task decomposition" icon="diagram-project">
    Turns a single sentence into a structured plan — stages, subtasks, dependencies, and acceptance criteria — without you lifting a finger.
  </Card>

  <Card title="Parallel execution" icon="layer-group">
    Runs multiple agents at the same time across stages and projects, so a half-day plan finishes in a fraction of the time.
  </Card>

  <Card title="Remote trigger" icon="signal-stream">
    Kick off work from anywhere. Manager doesn't need your laptop in front of you to start, continue, or finish a task.
  </Card>

  <Card title="Result delivery" icon="inbox-full">
    Finished work lands with the full plan, outputs, files, and a clear summary of what changed.
  </Card>

  <Card title="Reply in context" icon="reply">
    When Manager needs a decision, it asks inline. You answer in a thread without stopping the run — and the conversation stays attached to that task.
  </Card>

  <Card title="Trigger from Slack & Telegram" icon="comments">
    Connect Slack, Telegram, Feishu, or Discord. Send a message from your phone, and Manager picks it up, runs it, and replies where you asked.
  </Card>

  <Card title="Long-term memory" icon="brain">
    Manager remembers your projects, stack choices, conventions, and past decisions — so the next task starts from what it already knows about you.
  </Card>
</CardGroup>

***

### Task decomposition [#task-decomposition]

You describe the outcome; Manager figures out the steps. It identifies the goal, defines the stages (e.g. setup → core logic → UI → validation), splits each stage into executable subtasks, and decides which can run in parallel. You never need to plan, prioritize, or write a spec.

### Parallel execution [#parallel-execution]

Manager dispatches a dedicated worker per subtask and runs as many as possible at once. A multi-step build that would take a single agent hours often completes in minutes. Progress shows up live on the Kanban and in Pulse so you can watch — or just walk away.

### Remote trigger [#remote-trigger]

Manager keeps running whether you're watching or not. Start a task before bed, get coffee, hop on a flight — when you come back, completed work is waiting and anything that needed your input is queued up as a reply request.

### Result delivery [#result-delivery]

Every task ends with a clear handoff: the finished output, the files it touched, a summary of decisions, and a one-click path to the diffs or running app. Finished tasks gather in **To Review** so nothing slips past you.

### Reply [#reply]

When Manager hits a fork — a missing detail, a design choice, a permission — it asks you inline and keeps the rest of the work moving. Pending questions surface in a banner so you can clear them in one pass, and your reply is attached to the exact task that asked.

### Slack / Telegram message triggers [#slack--telegram-message-triggers]

Connect your messengers once and Manager treats them as inboxes:

* Slack, Telegram, Feishu, Discord — pair with the built-in guided setup
* A message from any connected channel becomes a tracked task
* Replies and final results route back to the same channel and thread

You can run an entire project from your phone without opening the app.

### Long-term memory [#long-term-memory]

Manager learns from past conversations. Import the last 7 days to 6 months of history, and Manager keeps your preferences — frameworks, naming, architecture decisions, ways of working — in mind for every new task. The more you use it, the less you have to re-explain.

***

## Tips to get the most out of Manager [#tips-to-get-the-most-out-of-manager]

* **State the outcome, not the steps.** "Ship a working signup flow with email + Google" beats a checklist.
* **Add acceptance criteria.** "Restart works on R key" is more useful than "make it good."
* **Use `@project-name`** to drop work into a specific project; skip it to let Manager create a fresh one.
* **Interrupt anytime.** "Actually, use TypeScript" mid-run is fine — Manager re-plans.
* **Schedule the things you'll want again.** Ask Manager to put a task on repeat and it shows up in Automations.

***

## Where to go next [#where-to-go-next]

<CardGroup cols="2">
  <Card title="Pulse" icon="chart-mixed" href="/verdent-manager/core-features/pulse">
    The live dashboard for every running task.
  </Card>

  <Card title="Automations" icon="clock" href="/verdent-manager/core-features/automation">
    Recurring and scheduled tasks in natural language.
  </Card>

  <Card title="Instant Messaging" icon="comments" href="/verdent-manager/core-features/instant-messaging">
    Trigger Manager from Slack, Telegram, Feishu, and Discord.
  </Card>
</CardGroup>