Cloudflare Search

English | 中文

An aggregated search API service based on Cloudflare Workers

Supports MCP (Model Context Protocol), giving AI assistants (OpenClaw, Claude Code, Codex, OpenCode) real-time web search capabilities

Features

• 🔍 Multi-engine Aggregation - Use multiple search engines at the same time (Google, Brave, DuckDuckGo, Bing)
• 🤖 AI Enhanced (MCP) - Native support for Model Context Protocol, one-click search tool integration for OpenClaw / Claude Code / Codex
• ⚡ Parallel Search - All search engines are requested concurrently for faster results
• 🛡️ Fault Tolerance - Failure of a single engine does not affect others; unresponsive engines are automatically marked
• ⏱️ Timeout Control - Configurable request timeout to avoid long waits
• 🔒 Token Authentication - Supports token auth to protect the service from abuse
• 🌍 CORS Support - Full cross-origin resource sharing support
• 🎨 Web Interface - Provides a clean search UI for easy testing
• ⚡ Zero-cost Operation - Cloudflare Workers free tier supports 100,000 requests per day

Page Preview

MCP Integration: Use in OpenClaw / Claude Code / AI Agents

With MCP (Model Context Protocol), AI assistants can directly call your search service and get real-time search results.

Installation and Configuration

1. Deploy the Service

First, follow the guide to Deploy Cloudflare Search

2. Add MCP Server Configuration

Edit your config file (configuration guide):

• OpenClaw: ~/.openclaw/openclaw.json
• Claude Code: ~/.claude/config.json / ~/.claude.json
• Claude Desktop macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Claude Desktop Windows: %APPDATA%\Claude\claude_desktop_config.json

{
	"mcpServers": {
		"cloudflare-search": {
			"command": "npx",
			"args": ["-y", "@yrobot/cf-search-mcp"],
			"env": {
				"CF_SEARCH_URL": "https://your-worker.workers.dev",
				"CF_SEARCH_TOKEN": "your-token-here"
			}
		}
	}
}

Environment Variables:

• CF_SEARCH_URL: Worker deployment URL (required)
• CF_SEARCH_TOKEN: Auth token (required if your Worker has TOKEN configured)

3. Verify Installation

• OpenClaw: Run openclaw gateway restart + openclaw mcp list and check that cloudflare-search appears
• Claude Code:
  • Run /mcp in Claude Code, and you should see the cloudflare-search tool.
  • Or run claude mcp list; seeing cloudflare-search: npx -y @yrobot/cf-search-mcp@latest - ✓ Connected means setup is successful

Installation Methods

Method 1: One-click Deployment (Recommended)

Click the "Deploy to Cloudflare Workers" button above and follow the prompts.

Method 2: Use Wrangler CLI

# 1. Install Wrangler
npm install -g wrangler

# 2. Login to Cloudflare
wrangler login

# 3. Clone the repository
git clone https://github.com/Yrobot/cloudflare-search.git
cd cloudflare-search

# 4. Deploy
wrangler deploy

Method 3: Use Cloudflare Dashboard

1. Sign in to Cloudflare Dashboard
2. Go to Workers & Pages
3. Click Create Application > Create Worker
4. Click Upload to upload your local code folder
   • Select the cloned cloudflare-search folder
   • Or manually copy worker.js, envs.js, utils/, and other files
5. Click Save and Deploy

Get Access URL

After deployment, you will get a Worker URL:

https://your-worker-name.your-subdomain.workers.dev

Note: The default domain may not be directly accessible in some regions. It is recommended to bind your own custom domain.

Usage

Method 1: Web Interface

Open your Worker URL directly and enter searc

... View full README on GitHub