BigQuery MCP server

@LucasHild

代码仓库

基本信息

分类

数据库

许可证

MIT

传输方式

stdio

作者

LucasHild

关于 BigQuery MCP server

A Model Context Protocol server that provides access to BigQuery

#bigquery#database-querying#llm-integration

BigQuery MCP server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "mcp-server-bigquery": {
      "command": "npx",
      "args": [
        "-y",
        "@smithery/cli",
        "install",
        "mcp-server-bigquery",
        "--client",
        "claude"
      ]
    }
  }
}

BigQuery MCP server 工具

Executes a SQL query using BigQuery dialect

Lists all tables in the BigQuery database

Describes the schema of a specific table

概览

What is BigQuery MCP server?

A Model Context Protocol server that provides access to BigQuery, enabling LLMs to inspect database schemas and execute queries. It is intended for developers using AI assistants to interact with BigQuery data.

How to use BigQuery MCP server?

Install via Smithery using npx -y @smithery/cli install mcp-server-bigquery --client claude or configure directly in client configs (Claude Desktop, Claude Code, Cursor). Required configuration: project ID (via --project argument or BIGQUERY_PROJECT env var) and location (via --location or BIGQUERY_LOCATION). Run with uvx mcp-server-bigquery.

Key features of BigQuery MCP server

Provides three tools: execute-query, list-tables, describe-table
Configurable via command line arguments or environment variables
Supports filtering by specific BigQuery datasets
Can use a service account key file or default credentials
Optional query timeout setting

Use cases of BigQuery MCP server

Enabling LLMs to discover BigQuery table schemas
Running SQL queries against BigQuery from an AI assistant
Interacting with BigQuery data without leaving the chat interface

常见问题

What tools does the BigQuery MCP server implement?

It implements three tools: `execute-query` (executes a SQL query using BigQuery dialect), `list-tables` (lists all tables in the database), and `describe-table` (describes the schema of a specific table).

What are the required configuration settings?

The GCP project ID and location are required. They can be provided via command line arguments (`--project`, `--location`) or environment variables (`BIGQUERY_PROJECT`, `BIGQUERY_LOCATION`).

Can I limit access to specific datasets?

Yes, use the `--dataset` argument (repeatable) or `BIGQUERY_DATASETS` environment variable with comma-separated dataset names. If not provided, all datasets in the project are considered.

How do I authenticate with BigQuery?

Either provide a path to a service account key file via `--key-file` or `BIGQUERY_KEY_FILE`, or omit it to use default credentials (e.g., from the environment or gcloud).

Is there a query timeout?

An optional `--timeout` argument (or `BIGQUERY_TIMEOUT` env var) sets the maximum time in seconds to wait for query completion. If not provided, queries wait indefinitely.