Tutorial

How to Connect MCP Search to Claude Desktop

Learn how to configure Scavio's MCP search server in Claude Desktop so Claude can search Google, Reddit, YouTube, Amazon, and Walmart during conversations.

Claude Desktop supports MCP (Model Context Protocol) servers that give Claude access to external tools during conversations. Scavio's MCP server exposes 11 search tools covering Google, Reddit, YouTube, Amazon, and Walmart. Once connected, Claude can search any of these platforms when answering your questions, grounding its responses in live data instead of training knowledge. This tutorial walks through the one-time configuration.

Prerequisites

  • Claude Desktop installed (macOS or Windows)
  • A Scavio API key from scavio.dev
  • Basic familiarity with JSON configuration files

Walkthrough

Step 1: Locate the Claude Desktop config file

Claude Desktop stores MCP server configuration in a JSON file. Open it in your text editor.

Bash
# macOS:
open ~/Library/Application\ Support/Claude/claude_desktop_config.json

# Windows:
notepad %APPDATA%\Claude\claude_desktop_config.json

# If the file does not exist, create it with: {}

Step 2: Add the Scavio MCP server

Add Scavio to the mcpServers object. Replace your_scavio_api_key with your actual key.

JSON
{
  "mcpServers": {
    "scavio": {
      "url": "https://mcp.scavio.dev/mcp",
      "headers": {
        "x-api-key": "your_scavio_api_key"
      }
    }
  }
}

Step 3: Restart Claude Desktop

Close and reopen Claude Desktop. In the tools panel, you should see 11 Scavio search tools available.

Bash
# After restart, verify by asking Claude:
# "What search tools do you have access to?"
# Claude should list google_search, reddit_search, youtube_search,
# amazon_search, walmart_search, and more.

Step 4: Test with a live search

Ask Claude a question that requires current information. Claude will call the appropriate search tool automatically.

Bash
# Example prompts that trigger search:
# "What are the top-rated wireless earbuds on Amazon right now?"
# "What is Reddit saying about the new MacBook Pro?"
# "Find me the latest tutorials on LangChain agents on YouTube"

Python Example

Python
# For programmatic MCP client (not Claude Desktop):
import subprocess, json

config = {
    'mcpServers': {
        'scavio': {
            'url': 'https://mcp.scavio.dev/mcp',
            'headers': {'x-api-key': 'your_scavio_api_key'}
        }
    }
}
print(json.dumps(config, indent=2))

JavaScript Example

JavaScript
// For Cursor or VS Code MCP configuration:
// Add to .cursor/mcp.json or .vscode/mcp.json:
const config = {
  mcpServers: {
    scavio: {
      url: 'https://mcp.scavio.dev/mcp',
      headers: { 'x-api-key': 'your_scavio_api_key' }
    }
  }
};
console.log(JSON.stringify(config, null, 2));

Expected Output

JSON
Claude Desktop with 11 Scavio search tools available. Claude can now search Google, Reddit, YouTube, Amazon, and Walmart during conversations.

Related Tutorials

    Frequently Asked Questions

    Most developers complete this tutorial in 15 to 30 minutes. You will need a Scavio API key (free tier works) and a working Python or JavaScript environment.

    Claude Desktop installed (macOS or Windows). A Scavio API key from scavio.dev. Basic familiarity with JSON configuration files. A Scavio API key gives you 500 free credits per month.

    Yes. The free tier includes 500 credits per month, which is more than enough to complete this tutorial and prototype a working solution.

    Scavio has a native LangChain package (langchain-scavio), an MCP server, and a plain REST API that works with any HTTP client. This tutorial uses the raw REST API, but you can adapt to your framework of choice.

    Start Building

    Learn how to configure Scavio's MCP search server in Claude Desktop so Claude can search Google, Reddit, YouTube, Amazon, and Walmart during conversations.