Kaggle

About Kaggle

Kaggle provides programmatic access to its platform through two tools: the kaggle CLI, a command-line interface for interacting with Kaggle via terminal commands or shell scripts, and kagglehub, a Python library for integrating Kaggle resources such as models, datasets, and competitions into Python machine learning workflows.

Authentication is available either through an OAuth flow using the kaggle CLI's login command, or via an API key generated from the account's API tokens settings page.

The API applies dynamic rate limiting across both the public API and the kaggle.com website. Requests that return an HTTP 429 "Too many requests" error can be retried after a pause; Kaggle also recommends reviewing client code for unintended loops or redundant calls, and reporting suspected platform-side rate-limiting bugs through the Product Feedback forum.

Key features

• Command-line interface (kaggle CLI) for interacting with Kaggle from a terminal or shell script
• kagglehub Python library for integrating Kaggle datasets, models, and competitions into ML workflows
• OAuth authentication via kaggle auth login
• API key generation from account API tokens settings page
• Dynamic rate limiting on both the public API and website calls

Frequently asked questions