cd product-researcher-mcp
pip install -r requirements.txt

cp .env.example .env
# Edit .env — set SEARCH_PROVIDER and the matching API key

{
  "mcpServers": {
    "product-researcher": {
      "command": "python",
      "args": ["/full/path/to/product-researcher-mcp/server.py"],
      "env": {
        "SEARCH_PROVIDER": "tavily",
        "TAVILY_API_KEY": "tvly-your-key-here"
      }
    }
  }
}

{
  "mcpServers": {
    "product-researcher": {
      "command": "python",
      "args": ["/full/path/to/product-researcher-mcp/server.py"],
      "env": {
        "SEARCH_PROVIDER": "brave",
        "BRAVE_API_KEY": "BSA-your-key-here"
      }
    }
  }
}

research Kling AI

compare Supabase vs Firebase vs PlanetScale

what's the pricing for Vercel

find alternatives to Midjourney for AI image generation

You ask Claude Code
    → Claude calls research_product tool
        → MCP server builds smart search queries
            → Hits your search API (Tavily/Brave/Serper)
                → Deduplicates & formats results
                    → Returns structured markdown brief
                        → Claude synthesizes the answer for you

product-researcher-mcp/
├── server.py          # The MCP server (all tools + search adapters)
├── requirements.txt   # Python dependencies
├── .env.example       # Config template
└── README.md          # You're here

async def _search_newprovider(query: str, max_results: int) -> List[Dict[str, Any]]:
    # Hit the API, return list of {"title": ..., "content": ..., "url": ...}
    pass