What is Gemini Lg Mcp

Gemini-lg-mcp is an AI-powered research assistant that utilizes Google’s Gemini AI models, Google Search, and LangGraph to perform comprehensive web research.

Use cases

Use cases include conducting thorough literature reviews, verifying facts for articles, analyzing market trends, gathering competitive insights, and exploring new areas of research.

How to use

To use gemini-lg-mcp, obtain a Gemini API key from the provided link, then add the necessary configuration to your .cursor/mcp.json or Claude Desktop mcp.json file, ensuring to include your API key.

Key features

Key features include smart search strategies that generate optimized queries, iterative research that automatically addresses knowledge gaps, citation tracking for source preservation, and structured output for organized data.

Where to use

Gemini-lg-mcp can be used in various fields such as academic research, fact-checking, market analysis, competitive intelligence, and staying updated on specific topics.

Gemini LangGraph Research MCP Server

🔬 AI-powered research assistant that performs comprehensive web research using Google’s Gemini AI models, Google Search and LangGraph.

What it does

• Smart search strategies - Generates multiple optimized search queries
• Iterative research - Follows up on knowledge gaps automatically using LangGraph
• Citation tracking - Preserves source URLs and references
• Structured output - Returns organized research data for easy use

Perfect for research, fact-checking, market analysis, competitive intelligence, and staying current on any topic.

Quick Start

1. Get a Gemini API Key

Get your free API key at: https://aistudio.google.com/app/prompts/new_chat

2. Add to Cursor / Claude Desktop

Add the following to your .cursor/mcp.json file or Claude Desktop mcp.json file:

{
  "mcpServers": {
    "gemini-research": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/albinjal/gemini-lg-mcp.git",
        "python",
        "-m",
        "src.server"
      ],
      "env": {
        "GEMINI_API_KEY": "your-api-key-here"
      }
    }
  }
}

Your client might not be able to find uvx. In that case run which uvx to find the path to uvx and add replace the command with the path to uvx:

{
  "mcpServers": {
    "gemini-research": {
      "command": "<output from which uvx>",
      "args": [
        "--from",
        "git+https://github.com/albinjal/gemini-lg-mcp.git",
        "python",
        "-m",
        "src.server"
      ],
      "env": {
        "GEMINI_API_KEY": "your-api-key-here"
      }
    }
  }
}

What you get back

• Search queries used - See the research strategy
• Research findings - Comprehensive results with citations
• Source URLs - All references preserved
• Knowledge gaps - What areas need more research
• Follow-up suggestions - Ideas for deeper investigation

Models Used

• Query generation: Gemini 2.0 Flash (fast, smart search strategies)
• Reflection: Gemini 2.5 Flash (identifies knowledge gaps)
• No final synthesis - Raw research data returned for your use

Installation

git clone https://github.com/albinjal/gemini-lg-mcp.git
cd gemini-lg-mcp
uv sync
export GEMINI_API_KEY="your-key"
uv run src/server.py

Why use this?

✅ Saves time - Automated multi-angle research
✅ Finds quality sources - Smart search with Google’s index
✅ Preserves citations - Never lose track of sources
✅ Customizable - Adjust depth and scope as needed

License

This project is licensed under the MIT License - see the LICENSE.md and the core logic is inspired by Gemini Fullstack LangGraph Quickstart

Built with LangGraph • Powered by Gemini • Designed for MCP