Claude Desktop
Claude Desktop Integration
Enable web scraping in Claude Desktop with CrawlForge MCP. Complete setup in under 5 minutes with a simple JSON configuration file.
Prerequisites
1
Install CrawlForge MCP Server
Install the CrawlForge MCP server package globally using npm.
Bash
npm install -g crawlforge-mcp-serverVerify installation: Run
crawlforge --version to confirm installation. The global install adds two server launchers—crawlforge-mcp and crawlforge-mcp-server—plus the crawlforge CLI.2
Get Your API Key
Sign up for a free account and generate your API key.
- Create a free account — Get 1,000 free credits—no credit card required. Sign Up Free
- Navigate to API Keys — Go to Dashboard → API Keys
- Generate a new key — Click "Create API Key" and copy the key (starts with
cf_)
Security: Never share your API key or commit it to version control. Store it securely.
3
Configure Claude Desktop
Add CrawlForge MCP to Claude Desktop's configuration file.
macOS
Open the configuration file at:
Bash
~/Library/Application Support/Claude/claude_desktop_config.jsonAdd the following configuration:
Json
{
"mcpServers": {
"crawlforge": {
"command": "npx",
"args": ["-y", "crawlforge-mcp-server"]
}
}
}Windows
Open the configuration file at:
Bash
%APPDATA%\Claude\claude_desktop_config.jsonAdd the following configuration:
Json
{
"mcpServers": {
"crawlforge": {
"command": "npx.cmd",
"args": ["-y", "crawlforge-mcp-server"]
}
}
}Linux
Open the configuration file at:
Bash
~/.config/Claude/claude_desktop_config.jsonAdd the following configuration:
Json
{
"mcpServers": {
"crawlforge": {
"command": "npx",
"args": ["-y", "crawlforge-mcp-server"]
}
}
}Where's the API key? This config has no inline key—the server reads the credentials that
npx crawlforge-setup stored in ~/.crawlforge/config.json. If you installed globally, you can also use "command": "crawlforge-mcp" (with no args) instead of npx—it resolves on your PATH and survives Node/nvm version switches.4
Restart and Test
Restart Claude Desktop and test the integration.
- Restart Claude Desktop — Fully quit and relaunch the application
- Open a new conversation — Start a fresh chat to load the MCP server
- Test with a simple prompt — Try: "Use CrawlForge to fetch https://example.com"
Example Prompts
"Fetch the content from https://example.com"
"Extract all links from https://example.com"
"Get metadata from https://example.com"
Available Tools
After setup, Claude will have access to all 23 CrawlForge MCP tools:
| Tool | Description | Credits |
|---|---|---|
| fetch_url | Fetch HTML from URLs | 1 credits |
| extract_text | Extract clean text | 1 credits |
| extract_links | Extract all links | 1 credits |
| extract_metadata | Get page metadata | 1 credits |
| scrape_structured | CSS selector extraction | 2 credits |
| batch_scrape | Parallel batch scraping | 1/URL |
| deep_research | AI-powered research | 10 credits |
| search_web | Google search integration | 5 credits |
See the full API reference for all 23 tools and their parameters.
Troubleshooting
Tools Not Showing Up?
- Verify the config file path is correct for your OS
- Check JSON syntax is valid (use a JSON validator)
- Ensure Node.js 18+ is installed:
node --version - Restart Claude Desktop completely (not just close window)
Authentication Errors?
- Verify API key starts with
cf_ - Check for spaces or quotes around the API key
- Generate a new API key from the dashboard
- Ensure key has not been deleted or expired
Rate Limit Issues?
Still having issues? Contact support at support@crawlforge.dev with your error message and config file (remove API key before sending).
Successfully integrated?
Explore all available tools and start building with AI-powered web scraping.