BigQuery MCP server

@MCP-Mirror

Repository

Basic information

About BigQuery MCP server

Mirror of

#bigquery #mcp-server #data-access

BigQuery MCP server config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "LucasHild_mcp-server-bigquery": {
      "command": "uv",
      "args": [
        "sync"
      ]
    }
  }
}

BigQuery MCP server tools

Executes a SQL query using BigQuery dialect

Lists all tables in the BigQuery database

Describes the schema of a specific table

Overview

What is BigQuery MCP server?

A Model Context Protocol server that gives LLMs access to Google BigQuery. It lets AI assistants inspect database schemas and execute SQL queries directly against BigQuery datasets.

How to use BigQuery MCP server?

Install via uvx mcp-server-bigquery and configure it in your MCP client (e.g., Claude Desktop) by passing --project (required), --location (required), and optionally --dataset (repeatable). The server runs over stdio and exposes three tools.

Key features of BigQuery MCP server

Execute arbitrary SQL queries using BigQuery dialect
List all tables in the configured project/datasets
Describe the schema of a specific table
Filter which datasets are available to the LLM
Published on PyPI for easy installation

Use cases of BigQuery MCP server

Let an LLM assistant answer natural-language questions by querying BigQuery
Automate schema discovery and documentation generation
Build a natural-language interface for business analysts to explore data
Integrate BigQuery querying into AI-powered coding or analysis workflows

Frequently asked questions

What tools does BigQuery MCP server provide?

It provides three tools: `execute-query`, `list-tables`, and `describe-table`.

How do I restrict which datasets the server can access?

Use the `--dataset` argument (repeatable) to specify one or more datasets. If omitted, all tables in the project are considered.

What are the required configuration arguments?

You must provide `--project` (the GCP project ID) and `--location` (the GCP location, e.g. `europe-west9`).

Is BigQuery MCP server published on PyPI?

Yes. It can be run via `uvx mcp-server-bigquery` without a local checkout.

How can I debug the server?

Use the MCP Inspector. Run `npx @modelcontextprotocol/inspector uv --directory <PATH> run mcp-server-bigquery` to get a debugging URL.

Comments

More Databases MCP servers

Multi Database MCP Server

FreePeak

A powerful multi-database server implementing the Model Context Protocol (MCP) to provide AI assistants with structured access to databases.

397

MCP Server for MySQL based on NodeJS

benborla

A Model Context Protocol server that provides read-only access to MySQL databases. This server enables LLMs to inspect database schemas and execute read-only queries.

1.9K