queryLABSCO SUMMARY
One tool, one job: current information from a search-capable model.
There is nothing to learn here — a single call takes a question and returns an answer, which is the right shape when all you want is to fill a gap in what your assistant knows. Everything rests on the underlying command-line tool being installed and configured, because the server is a thin passthrough to it.
What it is
A small server with a single tool: it passes your question to the Gemini command-line tool and returns the answer, giving an assistant a route to real-time information.
What you get
- query — send a question to an agent that can search real-time internet data, and get an answer back with current information
- A streaming HTTP transport, so clients connect to an endpoint instead of spawning a process
- Clean shutdown on interrupt and termination signals
Requirements
The Gemini command-line tool installed and configured on the machine running the server, since the server invokes it directly. Runs under Node from a checkout; clients connect over HTTP.
Setup effort
Build from source — git clone https://github.com/loming/gemini-mcp
USE CASES
01
Answering a question that postdates the model's knowledge
Send the question through and get back an answer drawn from a live search.
02
Checking a fact before you act on it
Ask for the current state of something rather than trusting a recalled figure.
03
Adding search to a client that has none
Point the client at the HTTP endpoint and it gains one search-capable tool.
1 tool
Listed in project docs. Your client may show a different list.
BEFORE YOU INSTALL
What to know first
Tools you get
1 tool named in the project's documentation.
Requires
- Node.js 18+
- gemini-cli installed and configured on the machine running the server
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
This one is built from source rather than installed with a single command.
Server setup
TERMINAL
git clone https://github.com/loming/gemini-mcp
cd gemini-mcp
npm install
npm run build
npm startThe server listens on http://localhost:13001/mcp (set PORT to change the port)
MCP Clients
Any MCP client — HTTP
ANY MCP CLIENT — HTTP · JSON
{
"mcpServers": {
"gemini-mcp": {
"url": "http://localhost:13001/mcp"
}
}
}Start the server first with npm start in your clone; it needs gemini-cli installed and configured on the same machine.
