Parallel Search MCP Server

A Model Context Protocol (MCP) server for Parallel Search API, built with ElysiaJS and deployed on Vercel.

Features

• 🔍 Web search with natural language objectives
• 🚀 Fast deployment on Vercel serverless
• 🔐 User-provided API keys (no cost to server owner)
• 📊 Ranked, LLM-optimized results with citations
• ⚙️ Customizable processors (base/pro)
• 🌐 Built with ElysiaJS and TypeScript

Deployment

Quick Deploy

1. Fork this repository
2. Connect to Vercel: https://vercel.com/new
3. Deploy!
4. Your MCP server will be available at: https://your-project.vercel.app/mcp

Local Development

# Install dependencies
bun install

# Run locally
bun run dev

# Test with MCP Inspector
npx @modelcontextprotocol/inspector@latest http://localhost:8000/mcp

Usage with Cursor

Add to your .cursor/mcp.json:

{
  "mcpServers": {
    "parallel-search": {
      "url": "https://your-deployment.vercel.app/mcp",
      "headers": {
        "x-api-key": "YOUR_PARALLEL_API_KEY"
      }
    }
  }
}

Note: Each user needs their own Parallel API key from https://parallel.ai

Tool: parallel_search

Parameters

• objective (optional): Natural language search goal
• search_queries (optional): List of specific queries (max 5)
• processor (optional): "base" (default) or "pro"
• max_results (optional): Limit number of results
• max_chars_per_result (optional): Control excerpt length (100-30000)
• source_policy (optional): Include/exclude specific domains
• api_key (optional): Your Parallel API key (or set via headers)

Example

In Cursor/Claude:

Search for "latest developments in quantum computing" using parallel_search

API Endpoints

• GET /health - Health check endpoint
• POST /mcp - MCP protocol endpoint

Development

# Install dependencies
bun install

# Run development server
bun run dev

# Build for production
bun run build

License

MIT