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

# MCP Overview

> Connect your AI agent to your Jedify data over the Model Context Protocol.

The Model Context Protocol (MCP) lets an AI agent — Claude Desktop, Cursor, or a hosted agent or app builder — work with your Jedify data through its semantic layer.

There are three modes:

* **Asker** *(default)* — ask questions in plain language and get data back.
* **Editor** — have your agent propose changes to your semantic layer for human review.
* **Builder** — have your agent build data apps and dashboards using Jedify's semantic functions.

Builder mode includes every Asker tool, plus additional tools and guidance for turning questions into reusable functions and wiring them into an app.

<Note>
  **Building a dashboard or app?** Read [Connect](/mcp/connect) → [Semantic functions](/mcp/builder/semantic-functions) → [Examples](/mcp/builder/examples). **Just asking questions?** See [Asker examples](/mcp/asker/use-cases).
</Note>

## Get started

<CardGroup cols={2}>
  <Card title="Modes" icon="toggle-on" href="/mcp/modes">
    Asker vs Builder, and how to switch.
  </Card>

  <Card title="Connect" icon="plug" href="/mcp/connect">
    Claude Desktop, Cursor, or a hosted agent.
  </Card>

  <Card title="Authentication" icon="key" href="/mcp/authentication">
    Sign in, or use an API key.
  </Card>

  <Card title="Troubleshooting" icon="bug" href="/mcp/troubleshooting">
    Common issues and fixes.
  </Card>
</CardGroup>