Gemini Google Web Search MCP
Provides Google Web Search functionality via the Gemini API, allowing for the retrieval of information with citations, source links, and grounding metadata.
Integrates with Google Cloud services and the Cloud AI Companion API to provide grounded web search results and support OAuth authentication.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Gemini Google Web Search MCPsearch for the latest news about the Model Context Protocol"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Gemini Google Web Search MCP Server
An MCP (Model Context Protocol) server that provides Google Web Search functionality using the Gemini API. This server is extracted from the official Gemini CLI's web search tool.
Features
Google Web Search via Gemini API
Citation support with source links
Grounding metadata for accurate information retrieval
Fully compatible with Claude Desktop and other MCP clients
Related MCP server: gemini-search
Prerequisites
Node.js 18+
Google API Key (Gemini API access)
Installation
Via npm (when published)
npm install -g gemini-google-web-search-mcpFrom source
git clone https://github.com/yourusername/gemini-google-web-search-mcp.git
cd gemini-google-web-search-mcp
npm install
npm run buildSetup
1. Get a Google API Key
Visit Google AI Studio to create your API key.
2. Configure Claude Desktop
Add the following to your Claude Desktop configuration file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Linux: ~/.config/Claude/claude_desktop_config.json
Option 1: API Key Authentication
{
"mcpServers": {
"google-web-search": {
"command": "node",
"args": ["/path/to/gemini-google-web-search-mcp/dist/index.js"],
"env": {
"GOOGLE_API_KEY": "your-api-key-here"
}
}
}
}Option 2: OAuth Authentication (No API Key Required)
{
"mcpServers": {
"google-web-search": {
"command": "node",
"args": ["/path/to/gemini-google-web-search-mcp/dist/index-codeassist.js"],
"env": {
"USE_OAUTH": "true"
}
}
}
}Usage
Once configured, the google_web_search tool will be available in Claude Desktop. You can use it by asking Claude to search for information:
"Search for the latest news about MCP protocol"
"Find information about TypeScript 5.0 features"
"Look up recent developments in AI safety"
Environment Variables
GOOGLE_API_KEYorGEMINI_API_KEY: Your Google API key (required)GEMINI_MODEL: Gemini model to use (default:gemini-2.0-flash-exp)USE_OAUTH: Set totrueto use OAuth authentication (experimental, not fully supported)GOOGLE_APPLICATION_CREDENTIALS: Path to service account credentials (optional)
Authentication Methods
1. API Key Authentication (Recommended)
The easiest way to authenticate is using an API key:
export GOOGLE_API_KEY=your-api-key-here2. OAuth Authentication with Code Assist API (Setup Required)
OAuth authentication uses Google's Code Assist API, which provides web search functionality with Google login.
Setup Steps:
Set OAuth credentials (if using custom OAuth app):
export OAUTH_CLIENT_ID=your-oauth-client-id export OAUTH_CLIENT_SECRET=your-oauth-client-secretRun the automated setup script:
./setup-gcp.shOr manually configure:
# Set your Google Cloud project ID export GOOGLE_CLOUD_PROJECT=your-project-id # Enable required APIs gcloud services enable cloudaicompanion.googleapis.com # Set up authentication gcloud auth application-default login \ --scopes=https://www.googleapis.com/auth/cloud-platform,https://www.googleapis.com/auth/generative-language.retrieverUse OAuth authentication:
export USE_OAUTH=true npm run dev:cli -- "your search query"
Requirements:
Google Cloud project with billing enabled
Cloud AI Companion API enabled
Gemini Code Assist subscription (free tier available)
Note: This uses the same authentication flow as Gemini CLI and accesses Google's Code Assist API for web search functionality.
Development
Building
npm run buildRunning in development
npm run devTesting
# Set your API key
export GOOGLE_API_KEY=your-api-key-here
# Run the test
npm testMCP Protocol Implementation
This server implements the Model Context Protocol with:
Tool:
google_web_searchInput:
query(string) - The search queryOutput: Search results with citations and source links
License
Apache 2.0 - This project is derived from Google's Gemini CLI which is licensed under Apache 2.0.
Credits
This MCP server is based on the web search functionality from Google's official Gemini CLI.
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
Troubleshooting
API Key Issues
Ensure your API key is valid and has access to Gemini API
Check that the environment variable is properly set
Connection Issues
Verify that the MCP server is properly configured in Claude Desktop
Check server logs for any error messages
Search Not Working
Ensure you have internet connectivity
Check if the Gemini API is accessible from your location
Verify that your API key has not exceeded rate limits
Available Tools
1 toolgoogle_web_searchB
Performs a web search using Google Search (via the Gemini API) and returns the results. This tool is useful for finding information on the internet based on a query.
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | The search query to find information on the web. |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It mentions the tool uses the Gemini API and returns results, but lacks details on behavioral traits such as rate limits, authentication needs, result format, or potential limitations. This leaves significant gaps for an agent to understand how to interact with it effectively.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise and front-loaded, with two sentences that directly state the tool's function and utility. There is no wasted text, but it could be slightly more structured (e.g., separating API details from usage). Overall, it earns its place efficiently.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the complexity of a web search tool, no annotations, and no output schema, the description is incomplete. It does not explain the return format, error handling, or operational constraints (e.g., API limits, latency), which are crucial for an agent to use the tool correctly in varied contexts.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 100% description coverage, with the 'query' parameter well-documented. The description adds minimal value beyond the schema by reiterating the query's purpose ('find information on the internet'), but does not provide additional semantics like query formatting tips or examples. Baseline 3 is appropriate given high schema coverage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Performs a web search using Google Search' and 'returns the results.' It specifies the action (search) and resource (web via Google), but lacks differentiation from siblings since there are none, which is acceptable but prevents a perfect score.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides implied usage by stating it's 'useful for finding information on the internet based on a query,' which suggests when to use it. However, it does not explicitly outline when not to use it or mention alternatives, and there are no siblings to compare against, so guidance is basic.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
TDQS
With only one tool, there is no possibility of ambiguity or overlap between tools. The tool's purpose is clearly defined as performing web searches, making it distinct by default.
Since there is only one tool, naming consistency is inherently perfect. The tool name 'google_web_search' follows a clear and descriptive pattern, though no comparison to other tools is possible.
A single tool for a web search server is too minimal for the apparent scope. While the tool covers the core function, typical search-related servers might include additional tools like advanced search options, filtering, or result management, making this feel thin and underdeveloped.
The server is severely incomplete for a web search domain. It lacks essential operations such as filtering results by date, language, or region, handling pagination, or providing related searches, which are common in search APIs and necessary for robust agent workflows.
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Live AI-native web search with citations. One tool for every MCP client. Flat per-request pricing.
MCP server unifying ERPs, CRMs, APIs and knowledge base for Claude, ChatGPT and Gemini.
Docs: https://docs.keenable.ai/mcp-server Keenable is a free, remote MCP server that gives agents access to the web index. Search the web with ranked results and date/site filters, then fetch any indexed page as clean markdown. Works out of the box with no account or API key.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceMCP server that integrates Google's Gemini 2.5 Pro with real-time Google Search grounding for current information retrieval.70MIT
- AlicenseNot gradedqualityDmaintenanceFree, unlimited web search MCP server using Gemini CLI's Google Search grounding to provide cited, live web results for any AI agent.1MIT
- AlicenseAqualityBmaintenanceMCP server for web search powered by Google AI Mode (Gemini). Enables any AI agent to search the web in real-time for free and without rate limits.2176MIT
- AlicenseAqualityBmaintenanceFree, unlimited web search MCP server powered by Google AI Mode (Gemini) without API key. Provides real-time search results for AI agents like Claude, Cursor, and Windsurf.2MIT
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/louxsye/google-web-search-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server