VSCode Automation

You run a specialized MCP server that lets AI agents automate and control VSCode. With this server, you can execute VSCode commands, inspect the UI and DOM, read editor content, and drive complex automation workflows directly from your MCP client.

How to use

You connect a compatible MCP client to the VSCode Automation MCP Server to perform a wide range of tasks. Start by launching the server from your preferred method, then invoke tools to run VSCode commands, simulate UI actions, inspect the DOM, read editor content, and capture diagnostics or screenshots. Use the available tools to automate development workflows, test extensions, and verify UI states.

How to install

Prerequisites you need on your machine: Node.js version 18 or newer and internet access for the initial setup. The server downloads a standalone VSCode instance and a matching ChromeDriver on first run to ensure consistent behavior without using your installed VSCode.

{
  "mcpServers": {
    "vscode-automation": {
      "command": "npx",
      "args": ["-y", "vscode-automation-mcp@latest"]
    }
  }
}

Local run and startup

You can also run the server directly if you have it installed globally. This starts the MCP server in the same way as the quick start example, using the CLI you installed.

vscode-automation-mcp

From source

If you prefer to build from source, clone the project, install dependencies, build, and start the server with these steps.

git clone https://github.com/sukarth/vscode-automation-mcp.git
cd vscode-automation-mcp
npm install
npm run build
npm start

Available tools

vscode_execute_command

Execute any VSCode command with optional arguments

vscode_list_commands

List available VSCode commands with optional filter

vscode_click_element

Click a UI element by CSS/XPath/accessibility selector

vscode_type_text

Type text into an input field or focused element

vscode_open_file

Open a file in the editor at a specific line/column

vscode_take_screenshot

Capture a screenshot of the VSCode window

vscode_get_element

Get properties and text of a UI element

vscode_open_webview

Open an extension webview panel

vscode_get_diagnostics

Get diagnostic messages from the Problems panel

vscode_get_dom

Get the full DOM structure of VSCode in tree, HTML, or JSON format

vscode_get_ui_structure

Get structure of UI regions like sidebar and editor

vscode_query_elements

Find all elements matching a CSS selector with attributes

vscode_get_accessibility_tree

Get the accessibility tree with ARIA roles and labels

vscode_get_editor_content

Get the current editor's text content

vscode_trigger_completion

Trigger IntelliSense/code completion

vscode_go_to_definition

Go to definition

vscode_trigger_hover

Trigger hover and get tooltip content

vscode_open_context_menu

Open context menu on an element

vscode_get_menu_items

Get items from visible menus

vscode_click_menu_item

Click a menu item by text

vscode_get_console_logs

Get captured console logs

vscode_get_output_channels

List available output channels

vscode_get_performance_metrics

Get performance metrics (memory, timing, DOM stats)

vscode_get_devtools_info

Get DevTools-style info