Connect to more than 20 Google Data Cloud services using powerful agentic tools and skills inside Visual Studio Code, Antigravity, Cursor, Claude Code, and beyond. No charge for Google Cloud users.
Features
Data Agent Kit provides a unified interface to interact with Cloud resources. Browse BigQuery schemas, edit Cloud Storage files, and monitor Spark clusters without having to switch back and forth between the Cloud Console and your IDE.
Simply describe your data-related task and have your agent reason and complete it for you. Data Agent Kit empowers your agent with tools and skills it needs to write notebooks, execute queries, deploy pipelines, and much more.
Data Agent Kit ships with opinionated agent skills from Google that span the entire data lifecycle—from writing optimized SQL and assessing the security posture of your storage buckets, to training and deploying machine learning models.
Installation
| Installation | |
|---|---|
| Environment | Instructions |
Visual Studio Code | Install from the Visual Studio marketplace |
Antigravity | Antigravity IDE
Antigravity CLI Run the following command in your terminal: agy plugin install https://github.com/gemini-cli-extensions/data-agent-kit-starter-pack Antigravity 2.0, and the Antigravity VSCode extension
|
Claude Code | Claude Code CLI Run the following command in your terminal: claude plugin install data-agent-kit-starter-pack@claude-plugins-official |
Codex | Codex CLI Run the following command in your terminal: codex plugin marketplace add gemini-cli-extensions/data-agent-kit-starter-pack codex plugin add dak@data-agent-kit-starter-pack-marketplace |
Google Cloud Shell / Workstation | Data Agent Kit comes preinstalled in both Google Cloud Shell and Google Cloud Workstation |
Cursor and other VSCode-based IDEs |
|
Installation
Antigravity
Antigravity IDE
Antigravity CLI
Run the following command in your terminal:
agy plugin install https://github.com/gemini-cli-extensions/data-agent-kit-starter-pack
Antigravity 2.0, and the Antigravity VSCode extension
Claude Code
Claude Code CLI
Run the following command in your terminal:
claude plugin install data-agent-kit-starter-pack@claude-plugins-official
Codex
Codex CLI
Run the following command in your terminal:
codex plugin marketplace add gemini-cli-extensions/data-agent-kit-starter-pack
codex plugin add dak@data-agent-kit-starter-pack-marketplace
Google Cloud Shell / Workstation
Data Agent Kit comes preinstalled in both Google Cloud Shell and Google Cloud Workstation
Cursor and other VSCode-based IDEs
How It Works
Data Agent Kit equips coding agents in your IDE or CLI with specialized MCP tools and skills. It allows your agent to translate natural language prompts directly into active SQL queries, pipeline code, or ML tasks, streamlining the entire data lifecycle without context switching.
Orchestrate, transform, and debug your pipelines
Build pipelines that turn source data into tables ready for downstream use, with quality checks to catch problems along the way. Develop transformations with dbt or Dataform and schedule workflows with Managed Service for Apache Airflow. Test changes before deployment, and when a run fails, work with your agent to trace the issue through execution logs and underlying code, then develop a fix.
Orchestrate, transform, and debug your pipelines
Build pipelines that turn source data into tables ready for downstream use, with quality checks to catch problems along the way. Develop transformations with dbt or Dataform and schedule workflows with Managed Service for Apache Airflow. Test changes before deployment, and when a run fails, work with your agent to trace the issue through execution logs and underlying code, then develop a fix.
Design features, train models, and automate inference
Develop and refine experiments in Jupyter notebooks with your agent using Python kernels or remote kernels on Managed Service for Apache Spark. Prepare features from BigQuery data and build predictive models using BigQuery Machine Learning. Connect the resulting work to scheduled batch inference that makes predictions available to downstream applications.
Design features, train models, and automate inference
Develop and refine experiments in Jupyter notebooks with your agent using Python kernels or remote kernels on Managed Service for Apache Spark. Prepare features from BigQuery data and build predictive models using BigQuery Machine Learning. Connect the resulting work to scheduled batch inference that makes predictions available to downstream applications.
Query databases, track lineage, and deliver business insights
Explore data across BigQuery and operational databases such as AlloyDB and Cloud SQL to answer business questions. Use Knowledge Catalog to find relevant datasets, with data quality metrics and lineage to guide your analysis. Work with your agent to develop queries and follow up on patterns in the results, then create visualizations or reusable data models for reporting and business decisions.
Query databases, track lineage, and deliver business insights
Explore data across BigQuery and operational databases such as AlloyDB and Cloud SQL to answer business questions. Use Knowledge Catalog to find relevant datasets, with data quality metrics and lineage to guide your analysis. Work with your agent to develop queries and follow up on patterns in the results, then create visualizations or reusable data models for reporting and business decisions.
Pricing
| Service | Price |
|---|---|
Data Agent Kit is available at no charge for users with Google Cloud accounts |
Data Agent Kit is available at no charge for users with Google Cloud accounts