webdev-mcp
# webdev-mcp
An MCP server that provides useful web development tools.
## Usage
### Cursor
- To install in a project, add the MCP server to your `.cursor/mcp.json`:
```json
{
"mcpServers": {
"webdev": {
"command": "npx",
"args": ["webdev-mcp"],
}
}
}
```
- To install globally, add this command to your Cursor settings:
```bash
npx webdev-mcp
```
### Windsurf
- Add the MCP server to your `~/.codeium/windsurf/mcp_config.json` file:
```json
{
"mcpServers": {
"webdev": {
"command": "npx",
"args": ["webdev-mcp"]
}
}
}
```
## Tools
Currently, the only 2 tools are `takeScreenshot` and `listScreens`. Your agent can use the list screens tool to get the screen id of the screen it wants to screenshot.
The tool will return the screenshot as a base64 encoded string.

## Tips
Make sure YOLO mode is on and MCP tools protection is off in your Cursor settings for the best experience. You might have to allow Cursor to record your screen on MacOS.
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: listScreens identifies available screens for capture, while takeScreenshot performs the actual screenshot operation on a specified screen. There is no overlap or ambiguity between these functions, making it easy for an agent to select the correct tool based on the task.
The naming is inconsistent: listScreens uses camelCase, while takeScreenshot uses camelCase but with a different verb style (list vs. take). Although both are camelCase, the lack of a uniform verb_noun pattern (e.g., list_screens, take_screenshot) and mixed verb conventions reduce predictability and readability.
With only 2 tools, the server feels too thin for the web development domain implied by the name 'webdev-mcp'. This limited set lacks essential operations such as screen recording, window management, or other common web development tasks, making it insufficient for comprehensive coverage.
The tool surface is significantly incomplete for a web development server. While listScreens and takeScreenshot cover basic screenshot functionality, there are obvious gaps like capturing specific windows, recording screen activity, or integrating with web development workflows (e.g., browser automation, code editing). This will likely cause agent failures in broader tasks.