Extract clean, structured data from any URL — directly from Claude, Cursor, or any MCP-compatible AI.
Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on Mac):
{
"mcpServers": {
"haunt": {
"command": "npx",
"args": ["-y", "@hauntapi/mcp-server"],
"env": {
"HAUNT_API_KEY": "your-api-key"
}
}
}
}Get a free API key at hauntapi.com — 100 requests/month, no credit card needed.
Extract clean data from any URL. Handles JavaScript rendering and Cloudflare bypass automatically.
Extract the article content from https://example.com/blog/post
Extract from multiple URLs at once (up to 10).
Extract the title and price from these product pages: [url1, url2, url3]
Pull metadata: title, description, Open Graph tags, Twitter cards, canonical URL.
Get the Open Graph metadata for https://example.com
- Free: 100 requests/month
- Pro: $0.01/request (pay only for what you use)
- API Docs: https://hauntapi.com/docs
- Get API Key: https://hauntapi.com
- GitHub: https://github.com/hauntapi/mcp-server