Gemini Search

Provides Gemini 2.5 Flash with Google Search Grounding for grounded, web-aware queries through an MCP interface.

javascript

0

GitHub Stars

javascript

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "kidapu-gemini-search-mcp": {
      "command": "node",
      "args": [
        "/path/to/Gemini-Search-MCP/build/index.js"
      ],
      "env": {
        "GEMINI_MODEL": "flash",
        "GEMINI_API_KEY": "YOUR_API_KEY",
        "GEMINI_WEB_SEARCH": "true",
        "GEMINI_THINKING_BUDGET": "-1"
      }
    }
  }
}

You can run Gemini 2.5 Flash with Google Search Grounding as an MCP server to add powerful search-grounded capabilities to your MCP client workflows. This MCP lets you leverage Gemini’s Flash model with integrated web grounding to answer questions and fetch information on demand within your own tooling.

How to use

You will use this MCP by starting it as a stdio MCP server and then connecting your MCP client to it. Before you start, provide your Google Gemini API key via the GEMINI_API_KEY environment variable. The server exposes the Gemini model options (for example, flash, flash-lite, and pro) and can enable web search grounding. Once running, your MCP client can request grounded answers, specify a model, and toggle web search as needed. Ensure your client passes the appropriate prompts and leverages the grounding feature to fetch up-to-date information when relevant.

How to install

Prerequisites: Node.js and npm (or a compatible Node runtime) must be installed on your system. You will also need a Google Gemini API key obtained from Google AI Studio.

Step 1: Clone the MCP server repository locally.

git clone https://github.com/kidapu/Gemini-Search-MCP.git
cd Gemini-Search-MCP

Step 2: Install dependencies.

pnpm install

Step 3: Build the project.

pnpm run build

Step 4: Create and configure your environment file. Copy the example and set GEMINI_API_KEY.

cp .env.example .env

Edit the .env file to set GEMINI_API_KEY to your Google Gemini API key.

Step 5: Run the MCP server locally as a stdio MCP server using Node to execute the built entry point.

node build/index.js

Registration with Claude Code

To register the MCP with Claude Code, add it as a standard stdio MCP server using your preferred client command.

claude mcp add gemini-search -s user -- node /path/to/Gemini-Search-MCP/build/index.js

Available tools

grounding_search

Gemini grounding-enabled search capability that integrates Google Search results into responses for up-to-date information.