CA Data Agent MCP Server Setup Guide

Pre-requisites

• Premium AI Subscription: Claude AI Pro/Team or ChatGPT Plus/Team subscription required. Free versions do not support custom MCP servers.
  Alternative: Limited experience available via Gemini CLI (Gemini App doesn't support custom MCP servers yet).
• BigQuery CA Agent: Have one or more BigQuery Conversational Analytics (CA) agents configured in your GCP project.
  Important: Make sure your CA Agent has a clear, descriptive description - the AI uses this to decide when to call this tool.

1Required Configuration

Enable Managed MCP Server:

Follow the guide to enable the managed server and confirm you have the appropriate IAM roles:

https://cloud.google.com/mcp/configure-mcp-ai-application

Create OAuth Client:

Follow the authentication guide to create an OAuth client:

https://cloud.google.com/mcp/authenticate-mcp#create-oauth-client

Configure the OAuth client according to your platform:

• Claude AI: Claude AI OAuth Setup
• ChatGPT: Use redirect URI: https://chatgpt.com/connector_platform_oauth_redirect

2Add BigQuery Managed MCP Server

Add the BigQuery managed MCP server to your AI platform:

1. In Claude AI: Settings > Connectors > Add custom connector
   In ChatGPT: Settings > Apps > Create app
2. Enter the Server URL:
   https://bigquery.googleapis.com/mcp
3. Configure OAuth using credentials from the JSON file downloaded in Step #1:
   • Client ID: client_id field
   • Client Secret: client_secret field
4. Save and connect to login with Google

Claude AI:

ChatGPT:

3Add CA Agent MCP Server

Add a CA Agent MCP server using the same process as Step #2, but with a different URL. You can repeat this step to add multiple agents for different domains.

Generate your CA Agent MCP URL:

Your CA Agent MCP URL:

You're All Set!

Once configured, you can ask your AI assistant questions about your data, and it will use the CA agent to query BigQuery and provide insights.

How it works:

1. The AI agent retrieves your CA Agent configuration and metadata using the CA MCP server
2. The AI agent uses the BQ Managed MCP server to query the relevant tables based on your question

Example prompts:

• "What were our top-selling products last quarter?"
• "Show me the trend of user signups over the past 6 months"
• "Compare revenue between regions"