Model Context Protocol low risk

Bigquery

This MCP server enables LLMs to query BigQuery data using natural language, providing a secure read-only interface for data analysis and exploration.

Connections & Capabilities

Connects To

GitHubGCP

Capabilities

read

Quickstart

Install

npx @smithery/cli install @ergut/mcp-bigquery-server --client claude

Config

{
  "mcpServers": {
    "bigquery": {
      "command": "node",
      "args": [
        "/path/to/your/clone/mcp-bigquery-server/dist/index.js",
        "--project-id",
        "your-project-id",
        "--location",
        "us-central1",
        "--key-file",
        "/path/to/service-account-key.json"
      ]
    }
  }
}

Exposed MCP Tools (2)

safe

BigQuery Query

Executes SQL queries against BigQuery datasets based on natural language input.

Queries are read-only and limited to 1GB of processing.

safe

Dataset Schema Exploration

Allows browsing and understanding the structure of BigQuery datasets, including tables and views.

Provides read-only access to dataset metadata.

Safety Assessment

This server is generally safe for querying BigQuery data due to its read-only nature and query limits. However, it's crucial to secure the service account and monitor query activity to prevent unauthorized access or data leakage.

Read-only access to BigQuery data
1GB query limit to prevent excessive resource consumption
Authentication via Google Cloud CLI or service account
Clear labeling of resource types (tables vs views)
No direct write or delete operations

Potential for data exposure if service account is misconfigured
Reliance on Google Cloud's security infrastructure
Limited control over the generated SQL queries
Dependency on Claude Desktop's security for MCP communication
Possible information leakage through query history