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

# Use Avala with AI

> Connect Avala to Claude, Claude Code, Cursor, VS Code, and ChatGPT with MCP

The [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) is an open standard that lets AI assistants call external tools and access data sources. Avala's MCP server exposes your datasets, projects, and exports as tools, so you can manage your annotation workflow through natural language in Claude, Claude Code, Cursor, VS Code, or ChatGPT.

<img src="https://mintcdn.com/avala/QI4wcFsXqwG8KGbR/images/integrations/mcp-architecture.svg?fit=max&auto=format&n=QI4wcFsXqwG8KGbR&q=85&s=ff648834b0b525bc504e9b3defbeb0bd" alt="MCP architecture: AI editor connects to Avala MCP server, which calls the Avala API and returns structured data" style={{width: '100%', maxHeight: '280px', objectFit: 'contain'}} className="hidden dark:block" width="420" height="100" data-path="images/integrations/mcp-architecture.svg" />

<img src="https://mintcdn.com/avala/QI4wcFsXqwG8KGbR/images/integrations/mcp-architecture-light.svg?fit=max&auto=format&n=QI4wcFsXqwG8KGbR&q=85&s=128d958e9f2fc9fe4aecb400478f2c2a" alt="MCP architecture: AI editor connects to Avala MCP server, which calls the Avala API and returns structured data" style={{width: '100%', maxHeight: '280px', objectFit: 'contain'}} className="dark:hidden" width="420" height="100" data-path="images/integrations/mcp-architecture-light.svg" />

## Quick Setup

Install the MCP server with any compatible AI client. You'll need **Node.js 18+** and an [Avala API key](/docs/api-reference/authentication).

For Claude Desktop, add to your config file (`~/Library/Application Support/Claude/claude_desktop_config.json` on macOS):

<Warning>
  Never commit config files containing real API keys to version control. Set `AVALA_API_KEY` as a system environment variable, or replace `"your-api-key"` with your actual key locally.
</Warning>

```json theme={null}
{
  "mcpServers": {
    "avala": {
      "command": "npx",
      "args": ["-y", "@avala-ai/mcp-server"],
      "env": {
        "AVALA_API_KEY": "your-api-key"
      }
    }
  }
}
```

Restart Claude Desktop. The Avala tools will appear in the tool picker.

<Tip>
  MCP runs read-only by default in the bundled server. To enable write/delete
  tools, set `AVALA_MCP_ENABLE_MUTATIONS=true` before launching the MCP server.
</Tip>

## What You Can Do

Once connected, you can ask your AI assistant to interact with Avala using natural language:

```
List all my datasets
```

```
Show the status of the "Pedestrian Detection" project
```

```
How many tasks are completed in "Lidar QA"?
```

```
Show export jobs for dataset ds_abc123
```

```
Show my workspace stats
```

The assistant calls the appropriate Avala MCP tool, processes the response, and presents the results. You can chain multiple requests in a single conversation — for example, list datasets, pick one, then trigger an export.

## Full Setup Guide

For VS Code, ChatGPT, and a complete list of available MCP tools, see the full [MCP Server setup guide](/docs/integrations/mcp-setup).

<CardGroup cols={2}>
  <Card title="" icon="plug" href="/docs/integrations/mcp-setup">
    <p style={{fontWeight: 600, fontSize: '18px', marginBottom: '4px', marginTop: '8px', color: 'inherit'}}>MCP Server Setup</p>
    <p style={{fontSize: '14px', marginTop: '0px', opacity: 0.6}}>Complete setup instructions for all supported AI clients and the full tool reference.</p>
  </Card>

  <Card title="" icon="key" href="/docs/api-reference/authentication">
    <p style={{fontWeight: 600, fontSize: '18px', marginBottom: '4px', marginTop: '8px', color: 'inherit'}}>API Authentication</p>
    <p style={{fontSize: '14px', marginTop: '0px', opacity: 0.6}}>Create and manage the API keys used by the MCP server.</p>
  </Card>
</CardGroup>