Content
# url-to-markdown
[中文文档](docs/README.zh-CN.md)
An MCP server that extracts web page content from URLs and converts it to clean Markdown.
## Features
- Extracts readable content from any URL using Mozilla Readability
- Converts HTML to clean Markdown via Turndown
- Built-in retry logic for transient failures
- SSRF protection blocks access to private networks
- Automatic fallback extraction when Readability fails
## Installation
```bash
pnpm install
pnpm build
```
## Usage
### With Claude Desktop
Add to your Claude MCP configuration:
```json
{
"mcpServers": {
"url-to-markdown": {
"command": "node",
"args": ["/path/to/url-to-markdown/build/index.js"]
}
}
}
```
### Tool: `extract`
Extracts content from a URL and returns Markdown.
**Parameters:**
- `url` (string, required): The URL to extract content from (must start with `http://` or `https://`)
**Returns:** Markdown content or an error message.
### Tool: `search`
Search WeChat public account articles.
**Parameters:**
- `query` (string, required): Search keyword
- `page` (number, optional): Page number, default 1
**Returns:** Search results list or an error message.
## Development
| Command | Description |
|---------|-------------|
| `pnpm dev` | Watch mode |
| `pnpm test` | Run tests |
| `pnpm build` | Build |
| `pnpm lint` | Lint |
| `pnpm format` | Format code |
## Contributing
See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
## License
This project is licensed under the [GNU Lesser General Public License v2.1](LICENSE).
