Skip to content

Getting Started with HoloViz MCP

This tutorial will guide you through installing and using HoloViz MCP for the first time. By the end, you'll have HoloViz MCP running and be able to ask your AI assistant questions about Panel components!

What You'll Learn

  • How to install HoloViz MCP
  • How to configure it with your AI assistant (VS Code, Claude Desktop, or Cursor)
  • How to use it to get help building Panel applications
  • How to verify everything is working correctly

Prerequisites

Before you begin, ensure you have:

  • Python 3.11 or newer installed on your system
  • uv package installer
  • An MCP-compatible AI assistant:
    • VS Code with GitHub Copilot extension
    • Claude Desktop application
    • Cursor IDE
    • Or any other MCP-compatible client

Step 1: Install HoloViz MCP

Open your terminal and install HoloViz MCP as a uv tool:

uv tool install holoviz-mcp[panel-extensions]

This command installs HoloViz MCP along with common Panel extensions, making them available for your AI assistant to reference.

What's happening? The uv tool manager creates an isolated environment for HoloViz MCP and installs all necessary dependencies.

Step 2: Create the Documentation Index

HoloViz MCP needs to index the HoloViz documentation to provide intelligent answers. Run:

uvx holoviz-mcp update

⏱️ This takes 5-10 minutes on first run as it downloads and indexes documentation from Panel, hvPlot, and other HoloViz libraries.

Tip: While this runs, continue to Step 3 to configure your IDE!

Step 3: Configure Your AI Assistant

Choose your AI assistant and follow the appropriate configuration:

VS Code + GitHub Copilot

  1. In VS Code, open the Command Palette (Ctrl+Shift+P or Cmd+Shift+P)
  2. Type "MCP: Edit Settings" and press Enter
  3. Add this configuration to your mcp.json file:
{
  "servers": {
    "holoviz": {
      "type": "stdio",
      "command": "uvx",
      "args": ["holoviz-mcp"]
    }
  },
  "inputs": []
}
  1. Save the file and restart VS Code

For remote development: Add this to your Remote mcp.json instead of User settings to ensure the server runs on the remote machine. For Workspace customization: Add this to your Workspace mcp.json to override the User or Remote settings.

Claude Desktop

  1. Locate your Claude Desktop configuration file:
  2. macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  3. Windows: %APPDATA%\Claude\claude_desktop_config.json

  4. Linux: ~/.config/Claude/claude_desktop_config.json

  5. Add this configuration:

{
  "mcpServers": {
    "holoviz": {
      "command": "uvx",
      "args": ["holoviz-mcp"]
    }
  }
}
  1. Save the file and restart Claude Desktop

Cursor

  1. Open Cursor Settings
  2. Navigate to Features → Model Context Protocol
  3. Click Add Server and enter:
{
  "name": "holoviz",
  "command": "uvx",
  "args": ["holoviz-mcp"]
}
  1. Save and restart Cursor

Step 4: Verify Installation

Let's verify that HoloViz MCP is working correctly!

Check Server Status

In VS Code, you can monitor the MCP server:

  1. Open the Output panel (View → Output)
  2. Select "MCP: holoviz" from the dropdown
  3. You should see log messages indicating the server is running

Test with Your AI Assistant

Open a chat with your AI assistant and try these questions:

  1. Component Discovery: 
    What Panel components are available for user input?

Note: In VS Code, you can prefix your prompt with #holoviz to explicitly request that the AI use the holoviz-mcp server tools for your query.

  1. Component Details: 
    What parameters does the Panel Button component accept?

If your AI assistant provides detailed, accurate answers with specific Panel component information, congratulations! HoloViz MCP is working correctly! 🎉

Step 5: Build Your First Dashboard

Now that everything is set up, let's build a simple dashboard:

  1. Create a new file called app.py

  2. Ask your AI assistant: 

    Create a Panel dashboard that displays a slider and shows the square of the slider's value. Use Panel best practices.

  3. Your AI assistant will provide code using HoloViz MCP's knowledge of Panel components!

What's Next?

Now that you have HoloViz MCP running, explore more:

Troubleshooting

Installation Issues

Problem: uv: command not found

Solution: Install uv by following the uv installation guide

Problem: Installation takes too long

Solution: This is normal! The first installation downloads many dependencies. Subsequent updates are much faster.

Configuration Issues

Problem: AI assistant doesn't recognize Panel components

Solution: 1. Check that the documentation index completed (Step 2) 2. Verify your configuration file is correct 3. Restart your IDE 4. Check the MCP server logs for errors

Server Issues

Problem: MCP server won't start

Solution: 1. Verify Python 3.11+ is installed: python --version 2. Check uv installation: uv --version 3. Try running the server directly: uvx holoviz-mcp 4. Check the server logs in VS Code's Output panel

For more help, see the Troubleshooting Guide or join the HoloViz Discord.

Summary

In this tutorial, you:

✅ Installed HoloViz MCP using uv ✅ Created the documentation index ✅ Configured your AI assistant ✅ Verified the installation ✅ Built your first Panel dashboard

You're now ready to use HoloViz MCP to accelerate your Panel development! Happy coding! 🚀