Kaggle Studio

Run Jupyter notebooks and Python scripts on Kaggle's cloud infrastructure directly from VS Code.

Setup

Authentication (Required)

You need valid Kaggle API credentials for all features to work properly. Choose one of the following methods:

Method 1: Kaggle API Token File (Recommended)

1. Generate API Token: Go to kaggle.com/settings/account → Scroll to API section → Click "Create New API Token"
2. Download kaggle.json: This downloads a kaggle.json file containing your credentials
3. Place credentials file: Move the kaggle.json file to ~/.kaggle/kaggle.json

   mkdir -p ~/.kaggle
   mv ~/Downloads/kaggle.json ~/.kaggle/kaggle.json
   chmod 600 ~/.kaggle/kaggle.json
   

Method 2: VS Code Sign In

1. Get API credentials from kaggle.com/settings → Create New API Token
2. Sign in: Command Palette → Kaggle: Sign In → Enter username and API key

⚠️ Important: If you experience authentication errors (401 Unauthorized), your API token may be expired. Generate a new token from your Kaggle account settings and update your credentials.

Commands

Authentication

• Kaggle: Sign In - Authenticate with Kaggle API
• Kaggle: Sign Out - Clear stored credentials
• Kaggle: Check API Status - Verify API connection

Notebook Operations

• Kaggle: Run Current Notebook - Push and run notebook on Kaggle
• Kaggle: Push & Run - Upload notebook without auto-download
• Kaggle: Download Outputs - Download results from last run

Project Setup

• Kaggle: Init Project - Create kaggle.yml and metadata files
• Kaggle: Link Notebook - Connect to existing Kaggle notebook
• Kaggle: Attach Dataset - Add dataset to project configuration

Datasets & Competitions

• Kaggle: Browse Dataset Files - View and download dataset files
• Kaggle: Download Dataset - Download entire dataset
• Kaggle: Submit to Competition - Submit predictions

1. Initialize project: Kaggle: Init Project (creates kaggle.yml and kernel-metadata.json)
2. Open notebook: Any .ipynb file
3. Run on Kaggle: Click 🚀 button or use Kaggle: Run Current Notebook
4. Get results: Outputs download automatically to .kaggle-outputs/

Tree Views

• My Notebooks - Your Kaggle notebooks with download/link options and search by language/type/competition
• Datasets - Browse and attach popular datasets with search functionality
• Competitions - View entered competitions, featured competitions, and search all categories
• Runs - Track notebook execution history

Configuration

Example kaggle.yml:

project: my-project
kernel_slug: username/my-project
code_file: notebook.ipynb
accelerator: gpu    # none | gpu | tpu
internet: true
datasets:
  - username/dataset-name

Support

• GitHub Issues
• Marketplace