Why

When Claude gets stuck — or just confidently wrong — the fastest unblocker is a different model with different training. This server makes that one tool call away:

• 🟦 Google Gemini for an alternative model perspective
• 🟪 Perplexity for fresh web-grounded analysis
• 🟧 Stack Overflow for accepted answers from real engineers

You stay in your Claude conversation. The second opinion comes to you.

Features

• 🌐 Multi-source synthesis — combines insights from three different sources into one answer
• 🔤 Automatic language detection from file extensions
• 📋 Code snippet extraction and clean formatting
• 📄 Markdown report generation for solutions
• 🧠 Git-aware context gathering — includes nearby files when relevant
• ⚡ Single tool call — no need to context-switch between chat apps

Configure in Claude Desktop

Edit your claude_desktop_config.json:

{
  "mcpServers": {
    "second-opinion": {
      "command": "node",
      "args": ["/absolute/path/to/second-opinion-mcp-server/build/index.js"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-api-key",
        "PERPLEXITY_API_KEY": "your-perplexity-api-key",
        "STACK_EXCHANGE_KEY": "your-stack-exchange-key"
      }
    }
  }
}

Restart Claude Desktop.

Required environment variables

Stack

• TypeScript + Node.js 18+
• @modelcontextprotocol/sdk
• @google/generative-ai (Gemini)
• Perplexity API
• Stack Exchange API

Related MCP servers

• 🔍 mcp-perplexity-server (⭐ 14) — Perplexity-only, lighter weight
• 🐙 github-meta-mcp-server — Natural-language GitHub repo management

Install

git clone https://github.com/joewilsonai/second-opinion-mcp-server
cd second-opinion-mcp-server
npm install
npm run build

Usage

Once installed, just ask Claude:

"Get a second opinion on this race condition I'm hitting in our worker pool."

Claude will call the tool, pass your code + context, and synthesize a response from all three sources.

License

MIT