Launching Gemini CLI extensions for Google Data Cloud

Sandeep Karmarkar — Wed, 10 Dec 2025 15:24:16 +0000

In June, Google introduced Gemini CLI, an open-source AI agent that brings the power of Gemini directly into your terminal. We’re excited to announce open-source Gemini CLI extensions for Google Data Cloud services.

Building applications and analyzing trends with services like Cloud SQL, AlloyDB and BigQuery has never been easier — all from your local development environment! Whether you’re just getting started or a seasoned developer, these extensions make common data interactions such as app development, deployment, operations, and data analytics more productive and easier. So, let’s jump right in!

Using a Data Cloud Gemini CLI extension

Before you get started, make sure you have enabled the APIs and configured the IAM permissions required to access specific services.

To retrieve the newest functionality, install the latest release of the Gemini CLI (v0.6.0):

npm install -g @google/gemini-cli@latest

Next, install the extension:

gemini extensions install https://github.com/gemini-cli-extensions/<EXTENSION>

Replace with the name of the service you want to use. For example, alloydb, cloud-sql-postgresql or bigquery-data-analytics.

Before starting the Gemini CLI, you’ll need to configure the extension to connect with your Google Cloud project by adding the required environment variables. The table below provides more information on the configuration required.

Extension Name	Description	Configuration
alloydb	Create resources and interact with AlloyDB for PostgreSQL databases and data.	Configuration
alloydb-observability	Monitor database performance and health for AlloyDB for PostgreSQL databases.	Configuration
bigquery-data-analytics	Discover and ask questions from BigQuery data.	Configuration
bigquery-conversational-analytics	Dive deeper, discover insights from BigQuery data using the built-in stateless agent offered by Conversational Analytics API.	Configuration
cloud-sql-mysql	Connect and interact with a Cloud SQL for MySQL database and data.	Configuration
cloud-sql-mysql-observability	Monitor database performance and health for Cloud SQL for MySQL databases.	Configuration
cloud-sql-postgresql	Connect and interact with a Cloud SQL for PostgreSQL database and data.	Configuration
cloud-sql-postgresql-observability	Monitor database performance and health for Cloud SQL for PostgreSQL databases.	Configuration
cloud-sql-sqlserver	Connect and interact with a Cloud SQL for SQL Server database and data.	Configuration
cloud-sql-sqlserver-observability	Monitor database performance and health for Cloud SQL for SQL Server databases.	Configuration
dataplex	Connect to Dataplex Universal Catalog to discover, manage, monitor, and govern data and AI artifacts across your data platform.	Configuration
spanner	Connect and interact with a Spanner database and data.	Configuration
mcp-toolbox	Load custom tools using MCP Toolbox for Databases.	Configuration

Now, you can start the Gemini CLI using command gemini. You can view the extensions installed with the command /extensions

You can list the MCP servers and tools included in the extension using command /mcp list

Using the Gemini CLI for Cloud SQL for PostgreSQL extension

The Cloud SQL for PostgreSQL extension lets you perform a number of actions. Some of the main ones are included below:

Create instance: Creates a new Cloud SQL instance for PostgreSQL (and also MySQL, or SQL Server)
List instances: Lists all Cloud SQL instances in a given project
Get instance: Retrieves information about a specific Cloud SQL instance
Create user: Creates a new user account within a specified Cloud SQL instance, supporting both standard and Cloud IAM users

Curious about how to put it in action? Like any good project, start with a solid written plan of what you are trying to do. Then, you can provide that project plan to the CLI as a series of prompts, and the agent will start provisioning the database and other resources:

After configuring the extension to connect to the new database, the agent can generate the required tables based on the approved plan. For easy testing, you can prompt the agent to add test data.

Now the agent can use the context it has to generate an API to make the data accessible.

As you can see, these extensions make it incredibly easy to start building with Google Cloud databases!

Using the BigQuery Analytics extensions

For your analytical needs, we are thrilled to give you a first look at the Gemini CLI extension for BigQuery Data Analytics. We are also excited to give access to the Conversational Analytics API through the BigQuery Conversational Analytics extension. This is the first step in our journey to bring the full power of BigQuery directly into your local coding environment, creating an integrated and unified workflow.

With this extension you can

Explore data: Use natural language to search for your tables.
Analyze: Ask business questions on the data and generate intelligent insights.
Dive deeper: Use conversational analytics APIs to dive deeper into the insights.
And extend: Use other tools or extensions to extend into advanced workflows like charting, reporting, code management, etc.

This initial release provides a comprehensive suite of tools to Gemini CLI:

Metadata tools: Discover and understand the BigQuery data landscape.
Query execution tool: Run any BigQuery query and get the results back, summarized to your console.
AI-powered forecasting: Leverage BigQuery's built-in AI.Forecast function for powerful time-series predictions directly from the command line.
Deeper data Insights: The“ask_data_insights” tool provides access to server-side BigQuery agent for richer data insights.
And more…

[Note: To use the conversational analytics extension you need to enable additional APIs. Refer to documentation for additional info.]

Here is an example journey with analytics extensions:

Explore and analyze your data , e.g.,

> find tables related to PyPi downloads

✦ I found the following tables related to PyPi downloads:

   * file_downloads: projects/bigquery-public-data/datasets/pypi/tables/file_downloads
   * distribution_metadata: projects/bigquery-public-data/datasets/pypi/tables/distribution_metadata

>  Using bigquery-public-data.pypi.file_downloads show me top 10 downloaded pypi packages this month 

✦ Here are the top 10 most downloaded PyPI packages this month:

   1. boto3: 685,007,866 downloads
   2. botocore: 531,034,851 downloads
   3. urllib3: 512,611,825 downloads
   4. requests: 464,595,806 downloads
   5. typing-extensions: 459,505,780 downloads
   6. certifi: 451,929,759 downloads
   7. charset-normalizer: 428,716,731 downloads
   8. idna: 409,262,986 downloads
   9. grpcio-status: 402,535,938 downloads
   10. aiobotocore: 399,650,559 downloads

Run deeper insights

Using “ask_data_insights” to trigger an agent on the BigQuery (Conversational analytics API) to answer your questions. The server side agent is smart enough to gather additional context about your data and offer deeper insights into your questions.

You can go further and generate charts and reports by mixing BigQuery data with your local tools. Here’s a prompt to try:

”using bigquery-public-data.pypi.file_downloads can you forecast downloads for the last four months of 2025 for package urllib3? Please plot a chart that includes actual downloads for the first 8 months, followed by the forecast for the last four months”

Get started today!