All Integrations
BigQuery
DatabasesQuery petabytes of data in Google BigQuery using natural language. Your agents can generate SQL, run analytics, and return structured insights from your data warehouse.
How to Connect
- 1Go to Settings and then Credentials.
- 2Add a BigQuery connection with your project ID and service account key.
- 3Create a Data Source and select the BigQuery credential.
- 4Add the Data Source to your agent.
Connection Methods
MCP Template (Recommended)
RecommendedInstall from the template marketplace. Select an existing credential and your agents can run dynamic SQL queries, list tables, and describe schemas.
- 1Create a database credential in Settings → Credentials
- 2Go to MCP Templates marketplace
- 3Find the BigQuery Query template
- 4Select your credential from the dropdown
- 5Install — your agent can now query the database
Data Source
Create a versioned data source with a fixed query template. Best for controlled, repeatable queries where the SQL is predefined.
- 1Create a database credential in Settings → Credentials
- 2Go to Data Sources and create a new data source
- 3Write your SQL query template with {{ parameter }} placeholders
- 4Add the data source as a tool to your agent
Security & SQL Injection Protection
- All database credentials are encrypted at rest with AES-256-GCM
- Agent-generated SQL is executed with read-only permissions when possible
- Query results are truncated at 50,000 rows / 100MB to prevent resource exhaustion
- Connection details never leave the server — credentials are resolved server-side
- Use parameterized queries with the params argument to prevent SQL injection
- LLM Guard content scanning can be enabled to detect prompt injection attempts
- SSRF protection prevents connections to private/internal network addresses