MCP File Preview Server
# MCP File Preview Server
A Model Context Protocol (MCP) server that provides HTML file preview and analysis capabilities. This server enables capturing full-page screenshots of local HTML files and analyzing their structure.
## Features
- **File Preview**: Capture full-page screenshots of HTML files with proper CSS styling
- **Content Analysis**: Analyze HTML structure (headings, paragraphs, images, links)
- **Local File Support**: Handle local file paths and resources
- **Screenshot Management**: Save screenshots to a dedicated directory
## Installation
1. Clone the repository:
```bash
git clone https://github.com/your-username/mcp-file-preview.git
cd mcp-file-preview
```
2. Install dependencies:
```bash
npm install
```
3. Build the project:
```bash
npm run build
```
## Configuration
Add the server to your Claude or Cline MCP settings:
### Claude Desktop App
Add to `~/Library/Application Support/Claude/claude_desktop_config.json`:
```json
{
"mcpServers": {
"file-preview": {
"command": "node",
"args": ["/path/to/mcp-file-preview/build/index.js"]
}
}
}
```
### Cline VSCode Extension
Add to VSCode's MCP settings:
```json
{
"mcpServers": {
"file-preview": {
"command": "node",
"args": ["/path/to/mcp-file-preview/build/index.js"]
}
}
}
```
## Usage
The server provides two main tools:
### preview_file
Captures a screenshot and returns HTML content:
```typescript
<use_mcp_tool>
<server_name>file-preview</server_name>
<tool_name>preview_file</tool_name>
<arguments>
{
"filePath": "/path/to/file.html",
"width": 1024, // optional
"height": 768 // optional
}
</arguments>
</use_mcp_tool>
```
Screenshots are saved to `screenshots/` directory in the project folder.
### analyze_content
Analyzes HTML structure:
```typescript
<use_mcp_tool>
<server_name>file-preview</server_name>
<tool_name>analyze_content</tool_name>
<arguments>
{
"filePath": "/path/to/file.html"
}
</arguments>
</use_mcp_tool>
```
Returns counts of:
- Headings
- Paragraphs
- Images
- Links
## Development
1. Install dependencies:
```bash
npm install @modelcontextprotocol/sdk puppeteer typescript @types/node @types/puppeteer
```
2. Make changes in `src/`
3. Build:
```bash
npm run build
```
4. Test locally:
```bash
npm run dev
```
## Implementation Details
The server uses the MCP SDK's Server class with proper initialization:
```typescript
this.server = new Server(
// Metadata object
{
name: 'file-preview-server',
version: '0.1.0'
},
// Options object with capabilities
{
capabilities: {
tools: {
preview_file: {
description: 'Preview local HTML file and capture screenshot',
inputSchema: {
// ... schema definition
}
}
}
}
}
);
```
Key points:
- Server constructor takes separate metadata and options objects
- Tools are declared in capabilities.tools
- Each tool needs a description and inputSchema
- Screenshots are saved to a local `screenshots/` directory
## Debugging
1. Use the MCP Inspector:
```bash
npx @modelcontextprotocol/inspector
```
2. Connect with:
- Transport Type: STDIO
- Command: node
- Arguments: /path/to/build/index.js
3. Check Claude OS logs if tools don't appear in the dropdown
## Contributing
Please read [CONTRIBUTING.md](CONTRIBUTING.md) for details on our code of conduct and the process for submitting pull requests.
## License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: analyze_content focuses on structural analysis of HTML content, while preview_file handles file preview and screenshot capture. There is no overlap in functionality, making it easy for an agent to select the appropriate tool without confusion.
Both tools follow a consistent verb_noun naming pattern (analyze_content and preview_file), with clear, descriptive names that align with their functions. There are no deviations or mixed conventions, ensuring predictability and readability.
With only two tools, the server feels thin for a 'File Preview Server' that might be expected to handle more file types or operations beyond HTML. While the tools are focused, the count is borderline too few for the apparent scope, limiting functionality.
The server is severely incomplete for file preview purposes, as it only supports HTML files with no coverage for other common formats (e.g., images, PDFs, text files). There are significant gaps in the tool surface that will likely cause agent failures when dealing with non-HTML content.