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

# Managed tasks

Use the BrowserPair dashboard when you want BrowserPair to coordinate planning, bounded execution, approvals, Skills, and task history for you.

<CardGroup cols={3}>
  <Card title="Choose the browser" icon="browser">Select the connected profile where the task is allowed to run.</Card>
  <Card title="Choose the websites" icon="globe">Define one or more explicit domains before execution begins.</Card>
  <Card title="Describe the outcome" icon="play">Tell BrowserPair what should happen; the task remains bounded by your settings.</Card>
</CardGroup>

## Requirements

Before a managed task can run:

* at least one browser must be connected;
* at least one allowed website must be entered; and
* a **Brain** provider must be configured.

<Note>A Reflex provider is optional. BrowserPair can still use bounded deterministic execution where available and escalate planning decisions to the Brain.</Note>

## Run a task

<Steps>
  <Step title="Open Tasks">Describe the desired outcome in the task composer.</Step>
  <Step title="Open Task settings">Choose the browser and allowed websites. Multiple explicit domains are supported.</Step>
  <Step title="Choose a Skill or auto-match">Leave **Skill** on auto-match or select a compatible enabled Skill explicitly.</Step>
  <Step title="Keep the defaults or override them">Account-level Brain, Reflex, trust threshold, observation policy, and execution limits are used unless you choose task-specific overrides.</Step>
  <Step title="Run and review">BrowserPair creates one durable task, surfaces approvals when needed, and records the final state in task history.</Step>
</Steps>

## Task settings

<CardGroup cols={2}>
  <Card title="Authority" icon="shield">Browser, allowed websites, and Skill selection define what the task may attempt.</Card>
  <Card title="Model overrides" icon="brain">Override Brain, fallbacks, Reflex, or trust only when a task needs different routing.</Card>
  <Card title="Execution limits" icon="gauge">Bound Reflex steps, Brain calls, Reflex calls, total steps, and maximum duration.</Card>

  <Card title="Runtime website grants" icon="globe">
    Review user-approved domain expansion that applies only to the current task.
  </Card>
</CardGroup>

<Warning>
  A Skill, provider, screenshot, or model response cannot add browser authority.
</Warning>

<CardGroup cols={3}>
  <Card title="Brain + Reflex" icon="brain" href="/brain-reflex">
    Configure model routing and observations.
  </Card>

  <Card title="Skills" icon="puzzle-piece" href="/skills-guide">
    Reuse workflow knowledge safely.
  </Card>

  <Card title="Permissions" icon="shield" href="/permissions">
    Understand approvals and scope.
  </Card>
</CardGroup>
