This is a server that lets your LLMs (like Claude) talk directly to your BigQuery data! Think of it as a friendly translator that sits between your AI assistant and your database, making sure they can chat securely and efficiently. No more writing SQL queries by hand - just chat naturally with your data! This server uses the Model Context Protocol (MCP), which is like a universal translator for AI
Add this skill
npx mdskills install ergut/mcp-bigquery-serverWell-documented BigQuery MCP server with clear setup, multiple auth options, and helpful examples
1# BigQuery MCP Server2[](https://smithery.ai/protocol/@ergut/mcp-bigquery-server)3<div align="center">4 <img src="assets/mcp-bigquery-server-logo.png" alt="BigQuery MCP Server Logo" width="400"/>5</div>67## What is this? π€89This is a server that lets your LLMs (like Claude) talk directly to your BigQuery data! Think of it as a friendly translator that sits between your AI assistant and your database, making sure they can chat securely and efficiently.1011### Quick Example12```text13You: "What were our top 10 customers last month?"14Claude: *queries your BigQuery database and gives you the answer in plain English*15```1617No more writing SQL queries by hand - just chat naturally with your data!1819## How Does It Work? π οΈ2021This server uses the Model Context Protocol (MCP), which is like a universal translator for AI-database communication. While MCP is designed to work with any AI model, right now it's available as a developer preview in Claude Desktop.2223Here's all you need to do:241. Set up authentication (see below)252. Add your project details to Claude Desktop's config file263. Start chatting with your BigQuery data naturally!2728### What Can It Do? π2930- Run SQL queries by just asking questions in plain English31- Access both tables and materialized views in your datasets32- Explore dataset schemas with clear labeling of resource types (tables vs views)33- Analyze data within safe limits (1GB query limit by default)34- Keep your data secure (read-only access)3536## Quick Start π3738### Prerequisites39- Node.js 14 or higher40- Google Cloud project with BigQuery enabled41- Either Google Cloud CLI installed or a service account key file42- Claude Desktop (currently the only supported LLM interface)4344### Option 1: Quick Install via Smithery (Recommended)45To install BigQuery MCP Server for Claude Desktop automatically via [Smithery](https://smithery.ai/protocol/@ergut/mcp-bigquery-server), run this command in your terminal:4647```bash48npx @smithery/cli install @ergut/mcp-bigquery-server --client claude49```50The installer will prompt you for:5152- Your Google Cloud project ID53- BigQuery location (defaults to us-central1)5455Once configured, Smithery will automatically update your Claude Desktop configuration and restart the application.5657### Option 2: Manual Setup58If you prefer manual configuration or need more control:59601. **Authenticate with Google Cloud** (choose one method):61 - Using Google Cloud CLI (great for development):62 ```bash63 gcloud auth application-default login64 ```65 - Using a service account (recommended for production):66 ```bash67 # Save your service account key file and use --key-file parameter68 # Remember to keep your service account key file secure and never commit it to version control69 ```70712. **Add to your Claude Desktop config**72 Add this to your `claude_desktop_config.json`:7374 - Basic configuration:75 ```json76 {77 "mcpServers": {78 "bigquery": {79 "command": "npx",80 "args": [81 "-y",82 "@ergut/mcp-bigquery-server",83 "--project-id",84 "your-project-id",85 "--location",86 "us-central1"87 ]88 }89 }90 }91 ```9293 - With service account:94 ```json95 {96 "mcpServers": {97 "bigquery": {98 "command": "npx",99 "args": [100 "-y",101 "@ergut/mcp-bigquery-server",102 "--project-id",103 "your-project-id",104 "--location",105 "us-central1",106 "--key-file",107 "/path/to/service-account-key.json"108 ]109 }110 }111 }112 ```1131141153. **Start chatting!**116 Open Claude Desktop and start asking questions about your data.117118### Command Line Arguments119120The server accepts the following arguments:121- `--project-id`: (Required) Your Google Cloud project ID122- `--location`: (Optional) BigQuery location, defaults to 'us-central1'123- `--key-file`: (Optional) Path to service account key JSON file124125Example using service account:126```bash127npx @ergut/mcp-bigquery-server --project-id your-project-id --location europe-west1 --key-file /path/to/key.json128```129130### Permissions Needed131132You'll need one of these:133- `roles/bigquery.user` (recommended)134- OR both:135 - `roles/bigquery.dataViewer`136 - `roles/bigquery.jobUser`137138## Developer Setup (Optional) π§139140Want to customize or contribute? Here's how to set it up locally:141142```bash143# Clone and install144git clone https://github.com/ergut/mcp-bigquery-server145cd mcp-bigquery-server146npm install147148# Build149npm run build150```151152Then update your Claude Desktop config to point to your local build:153```json154{155 "mcpServers": {156 "bigquery": {157 "command": "node",158 "args": [159 "/path/to/your/clone/mcp-bigquery-server/dist/index.js",160 "--project-id",161 "your-project-id",162 "--location",163 "us-central1",164 "--key-file",165 "/path/to/service-account-key.json"166 ]167 }168 }169}170```171172## Current Limitations β οΈ173174- MCP support is currently only available in Claude Desktop (developer preview)175- Connections are limited to local MCP servers running on the same machine176- Queries are read-only with a 1GB processing limit177- While both tables and views are supported, some complex view types might have limitations178179## Support & Resources π¬180181- π [Report issues](https://github.com/ergut/mcp-bigquery-server/issues)182- π‘ [Feature requests](https://github.com/ergut/mcp-bigquery-server/issues)183- π [Documentation](https://github.com/ergut/mcp-bigquery-server)184185## License π186187MIT License - See [LICENSE](LICENSE) file for details.188189## Author βοΈ190191Salih ErgΓΌt192193## Sponsorship194195This project is proudly sponsored by:196197<div align="center">198 <a href="https://www.oredata.com">199 <img src="assets/oredata-logo-nobg.png" alt="OREDATA" width="300"/>200 </a>201</div>202203## Version History π204205See [CHANGELOG.md](CHANGELOG.md) for updates and version history.
Full transparency β inspect the skill content before installing.