Jina MCP Advanced

Pure MCP Server for Jina.AI Advanced web scraping capabilities. This server provides access to Jina's powerful web content extraction API through the Model Context Protocol.

Features

• 🔍 Advanced web scraping with Jina Reader
• 📄 Multiple output formats (markdown, HTML, text, screenshots)
• 🧠 Support for JavaScript-heavy sites with browser engine
• 🎯 CSS selector targeting for specific content
• 🖼️ Image handling and captioning
• ⚡ Fast and reliable content extraction

Installation

NPX (Recommended)

# No installation needed - run directly
npx @ildunari/jina-mcp-advanced

Global Installation

npm install -g @ildunari/jina-mcp-advanced
jina-mcp

Setup

1. Get Jina API Key

Get your free API key from Jina AI.

2. Claude Desktop Configuration

Add this to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "jina-advanced": {
      "command": "npx",
      "args": ["@ildunari/jina-mcp-advanced"],
      "env": {
        "JINA_API_KEY": "your-jina-api-key-here"
      }
    }
  }
}

Or if installed globally:

{
  "mcpServers": {
    "jina-advanced": {
      "command": "jina-mcp",
      "env": {
        "JINA_API_KEY": "your-jina-api-key-here"
      }
    }
  }
}

3. Restart Claude Desktop

After updating the configuration, restart Claude Desktop to load the new MCP server.

Available Tools

read_webpage

Fetch and extract content from any URL using Jina Reader Advanced.

Parameters:

• url (required) - The URL to scrape
• return_format - Output format: markdown, html, text, screenshot, pageshot (default: markdown)
• engine - Rendering engine: default or browser (use browser for JS-heavy sites)
• retain_images - Whether to include images (default: false)
• summarize_links - Include link summaries (default: true)
• summarize_images - Include image summaries (default: false)
• image_caption - Generate image captions (default: false)
• use_readerlm - Use ReaderLM for enhanced understanding (default: false)
• bypass_cache - Force fresh content fetch (default: false)
• target_selector - CSS selector for specific content
• wait_for_selector - Wait for element before scraping
• exclude_selector - Elements to exclude from scraping
• timeout - Request timeout in seconds

Usage Examples

After setup, you can use these commands in Claude:

Read the content from https://example.com and summarize it

Extract the main article from https://news-site.com using CSS selector ".article-content"

Take a screenshot of https://dashboard.example.com using browser engine

Environment Variables

• JINA_API_KEY - Your Jina AI API key (required)

Troubleshooting

MCP Server Not Working

1. Check configuration: Ensure your claude_desktop_config.json is valid JSON
2. Verify API key: Make sure your JINA_API_KEY is correct
3. Restart Claude: Restart Claude Desktop after configuration changes
4. Check logs: Look at Claude Desktop's logs for error messages

API Key Issues

• Get your API key from Jina AI Dashboard
• Make sure the key has sufficient credits/permissions
• Verify the key is correctly set in the environment variables

License

MIT

Support

For issues and feature requests, please visit the GitHub repository.