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

# Getting started

> Download Lace, grant macOS permissions, sign in, and capture your first comment.

## Install

Download Lace from [inlace.co](https://inlace.co) and drag it into your Applications folder. Open the app from Applications or Spotlight.

## Grant permissions

Lace needs four macOS permissions for full functionality:

| Permission           | Why                                                                          |
| -------------------- | ---------------------------------------------------------------------------- |
| **Screen recording** | Captures a screenshot of your active window                                  |
| **Accessibility**    | Reads text, labels, roles, and interactive state from the accessibility tree |
| **Microphone**       | Transcribes voice input for dictation in chat                                |
| **Notifications**    | Alerts you when teammates mention you or reply                               |

macOS prompts for screen recording and accessibility on first launch, microphone the first time you use voice dictation, and notifications when Lace first sends one. You can manage all four in **System Settings**. See [Permissions](/reference/permissions) for details.

## Sign in

Open Lace and sign in with your account. Your workspace is scoped to your organization.

## Start a conversation

Press **⌘L** to show or hide Lace. Type a message to start a new chat. To bring your screen into the conversation, press **Start session** in the app header (or **⌘S** while Lace is focused), and Lace uses what you're looking at as context. Use the model picker in the composer to choose which AI model responds.

<Tip>
  Chats can belong to a project or exist on their own. Use projects to group related conversations around a product area.
</Tip>

## Your first capture

<Steps>
  <Step title="Work like you normally do">
    Open the tools you use for a real task and start a **session** from the app header (or **⌘S**). Lace follows along, keeps up with what you're looking at, and records a [session flow](/concepts/sessions) agents can query later.
  </Step>

  <Step title="Mark what matters">
    Most moments don't need a note. When you decide something your team or your agent should understand, switch into Review (**⌘\\**, or the mode button on the pill) and pin it where it happened. Attach it as context for a comment.
  </Step>

  <Step title="Resolve it">
    Resolving promotes the comment to a **decision**. Lace attaches the screenshot, the spot on screen, and the discussion on its own.
  </Step>

  <Step title="Replay it">
    Connect your agent through [MCP](/reference/mcp). It queries your comments and decisions and executes with the full context behind each one.
  </Step>
</Steps>

## Attach files

Open the **+** menu in the composer and choose **Add files** to attach a document to any chat. Lace indexes the file and cites specific sections inline in its responses.

Supported formats: **PDF**, **CSV**, **Markdown**, **Word (.docx)**, and **plain text (.txt)**. Limits: 50 MB per file for PDF and CSV, 25 MB for Word, Markdown, and plain text, 10 files per chat.

## What to try first

<CardGroup cols={2}>
  <Card title="Ask about your screen" icon="desktop" href="/concepts/capture">
    Open any app and ask Lace what it sees.
  </Card>

  <Card title="Write a comment" icon="lightbulb" href="/concepts/reviews">
    Attach screen context to a comment. Resolve it into a decision.
  </Card>

  <Card title="Hand knowledge to your agent" icon="plug" href="/reference/mcp">
    Agents execute from what your team captured.
  </Card>

  <Card title="Publish for review" icon="globe" href="/reference/mcp">
    Your agent presents its work on a page, and your team's feedback is captured where they give it.
  </Card>
</CardGroup>