--- title: "BigQuery MCP Server: Claude and ChatGPT Setup | SourceMedium" description: Google's BigQuery MCP server at bigquery.googleapis.com/mcp, generally available since April 2026, lists datasets and runs SQL under IAM roles. url: "https://sourcemedium.com/mcp-directory/bigquery/" type: static generatedAt: "2026-09-29T01:21:01.000Z" --- [MCP Directory](https://sourcemedium.com/mcp-directory/) Data infrastructure Official # BigQuery MCP server Remote endpoint bigquery.googleapis.com/mcp. Checked September 2026. Google's hosted BigQuery server lists a project's datasets and tables, reads their metadata, and runs SQL under the caller's IAM roles over whatever a brand has loaded, such as GA4 exports, Shopify orders, or ad spend from scheduled transfers. Its execute_sql tool also runs INSERT, UPDATE, and DELETE unless an admin denies it by IAM policy. ## BigQuery MCP options Every fact in a row comes from the publisher's own documentation. | Option | Audience | Reads or changes | Sign-in | Setup | | --- | --- | --- | --- | --- | | BigQuery MCP server `https://bigquery.googleapis.com/mcp` Google | Merchants | Hosted by Google, it lists datasets and tables, reads their metadata and runs SQL under the caller's IAM roles; execute_sql_readonly allows SELECT only, while execute_sql can also run INSERT, UPDATE, DELETE, and CREATE statements unless an admin denies that tool with an IAM policy. Can write | OAuth sign-in | [Setup guide →](https://docs.cloud.google.com/bigquery/docs/use-bigquery-mcp) | | MCP Toolbox for Databases Google [Registered as io.github.googleapis/mcp-toolbox](https://registry.modelcontextprotocol.io/?q=io.github.googleapis%2Fmcp-toolbox) | Developers | An open-source server run locally with its prebuilt BigQuery toolset: lists datasets and tables, searches the catalog in natural language, executes SQL statements, and calls BigQuery forecasting, contribution analysis, and data insights tools, signed in with Application Default Credentials. Can write | See guide | [Setup guide →](https://docs.cloud.google.com/bigquery/docs/pre-built-tools-with-mcp-toolbox) | | BigQuery Data Transfer Service MCP server `https://bigquerydatatransfer.googleapis.com/mcp` Google | Developers | Lists data sources, transfer configurations, runs, and logs, and can create, update, delete, and manually start scheduled transfers such as a Google Ads load into BigQuery. Can write | OAuth sign-in | [Setup guide →](https://docs.cloud.google.com/bigquery/docs/reference/datatransfer/mcp) | ## Connect BigQuery to your assistant Each card lists the servers whose setup guide covers that assistant. ### Connect BigQuery to ChatGPT or Gemini - [BigQuery MCP server →](https://docs.cloud.google.com/bigquery/docs/use-bigquery-mcp) Guide covers ChatGPT, Gemini CLI - [BigQuery Data Transfer Service MCP server →](https://docs.cloud.google.com/bigquery/docs/reference/datatransfer/mcp) Guide covers ChatGPT, Gemini CLI ### Connect BigQuery to Claude - [BigQuery MCP server →](https://docs.cloud.google.com/bigquery/docs/use-bigquery-mcp) [Add from Claude's connector directory](https://claude.com/marketplace/connectors/bigquery) Claude Code ``` claude mcp add --transport http bigquery https://bigquery.googleapis.com/mcp ``` - [MCP Toolbox for Databases →](https://docs.cloud.google.com/bigquery/docs/pre-built-tools-with-mcp-toolbox) Guide also covers Claude Code - [BigQuery Data Transfer Service MCP server →](https://docs.cloud.google.com/bigquery/docs/reference/datatransfer/mcp) Claude Code ``` claude mcp add --transport http bigquery https://bigquerydatatransfer.googleapis.com/mcp ``` ### Connect BigQuery to Cursor - [BigQuery MCP server →](https://docs.cloud.google.com/bigquery/docs/use-bigquery-mcp) .cursor/mcp.json ``` {"mcpServers":{"bigquery":{"url":"https://bigquery.googleapis.com/mcp"}}} ``` - [MCP Toolbox for Databases →](https://docs.cloud.google.com/bigquery/docs/pre-built-tools-with-mcp-toolbox) - [BigQuery Data Transfer Service MCP server →](https://docs.cloud.google.com/bigquery/docs/reference/datatransfer/mcp) .cursor/mcp.json ``` {"mcpServers":{"bigquery":{"url":"https://bigquerydatatransfer.googleapis.com/mcp"}}} ``` ### Connect BigQuery to VS Code - [BigQuery MCP server →](https://docs.cloud.google.com/bigquery/docs/use-bigquery-mcp) .vscode/mcp.json ``` {"servers":{"bigquery":{"type":"http","url":"https://bigquery.googleapis.com/mcp"}}} ``` - [MCP Toolbox for Databases →](https://docs.cloud.google.com/bigquery/docs/pre-built-tools-with-mcp-toolbox) - [BigQuery Data Transfer Service MCP server →](https://docs.cloud.google.com/bigquery/docs/reference/datatransfer/mcp) .vscode/mcp.json ``` {"servers":{"bigquery":{"type":"http","url":"https://bigquerydatatransfer.googleapis.com/mcp"}}} ``` BigQuery's setup guides also name Antigravity, Windsurf, and Cline. ## Tools each server exposes Listed as each publisher documents them. A write tool can change something in the account. ### BigQuery MCP server 9 tools: 7 read, 2 write Read - `list_dataset_ids` - `get_dataset_info` - `list_table_ids` - `get_table_info` - `execute_sql_readonly` - `get_query_results` - `get_job` Write - `execute_sql` - `cancel_job` ### MCP Toolbox for Databases 9 tools: 8 read, 1 write Read - `analyze_contribution` - `ask_data_insights` - `forecast` - `get_dataset_info` - `get_table_info` - `list_dataset_ids` - `list_table_ids` - `search_catalog` Write - `execute_sql` ### BigQuery Data Transfer Service MCP server 13 tools: 8 read, 5 write Read - `list_data_sources` - `get_data_source` - `get_transfer_config` - `list_transfer_configs` - `list_transfer_runs` - `get_transfer_run` - `list_transfer_logs` - `check_valid_creds` Write - `create_transfer_config` - `update_transfer_config` - `delete_transfer_config` - `start_manual_transfer_runs` - `delete_transfer_run` ## Other ways to reach BigQuery data - [BigQuery API reference →](https://docs.cloud.google.com/bigquery/docs/reference/rest) - [Conversational analytics in BigQuery →](https://docs.cloud.google.com/bigquery/docs/conversational-analytics) ## One MCP server per tool still can't answer a cross-channel question. Each server brings its own sign-in, its own definition of revenue, and data your assistant can't join to the rest. SourceMedium connects 63 commerce and marketing sources, keeps one set of maintained definitions, and gives ChatGPT, Claude, and other assistants a single connection to all of it. BigQuery is not a SourceMedium source today. ## More in Data infrastructure - [Databricks MCP server →](https://sourcemedium.com/mcp-directory/databricks/) - [dbt Cloud MCP server →](https://sourcemedium.com/mcp-directory/dbt-cloud/) - [Fivetran MCP server →](https://sourcemedium.com/mcp-directory/fivetran/)