Kaggle/kaggle-cli

Official Kaggle CLI

What it solves

It provides a command-line interface (CLI) for users to interact with the Kaggle platform without needing to use a web browser, enabling the automation of data science workflows.

How it works

The tool acts as a bridge between the user's local terminal and the Kaggle website, allowing users to execute commands to manage their assets and participate in competitions via the kaggle package.

Who it’s for

Data scientists and machine learning engineers who want to automate the downloading of datasets, submission of competition entries, and management of notebooks and models.

Highlights

  • Competition Management: List, download data for, and submit entries to competitions.
  • Dataset Control: Create, update, download, or delete datasets.
  • Model Management: Handle models and their variations.
  • Kernel Operations: Run, download, or delete notebooks (kernels) and their outputs.
  • Forum Access: Browse and read discussion forums directly from the terminal.

Related

  • Project
    Kaggle/kagglehub

    A Python library for programmatically downloading, uploading, and loading Kaggle datasets, models, and competition resources into data science frameworks.

  • Project
    NVIDIA/nvidia-kaggle

    A Python‑based Agent Skill that lets LLM agents interact with Kaggle: fetch competition info, summarize top write‑ups, search discussions, download and reproduce notebooks, submit solutions, and upload datasets—all via natural‑language commands.

  • Project
    aandrew-me/tgpt

    A cross-platform CLI tool that allows users to interact with AI models directly from their terminal.

  • Project
    anthropics/anthropic-cli

    The `ant` CLI is the official command‑line interface for Anthropic’s Claude Platform. It lets you call Claude’s API from a terminal, manage agents, skills, environments, vaults, and other resources, and treat those resources as code via the `ant apply` workflow. Install via Homebrew or Go, authenticate with an API key or Google Cloud, and use a resource‑based command syntax (`ant <resource>[:<subresource>] <command>`). A lockfile records remote IDs for repeatable CI/CD deployments. MIT‑licensed and cross‑platform.

  • Dispatch
    Hugging Face and Kaggle Integration for Model Access

    Hugging Face and Kaggle have launched an integration that improves the discoverability and usability of Hugging Face models directly within Kaggle notebooks and model pages.