Skip to content

Croissant Toolkit

Croissant Toolkit is an intelligent suite of tools designed to work seamlessly with the MLCommons Croissant metadata format for machine learning datasets.

Code Video Cerebral Valley

Core trick

Sends dataset files to Gemini 3 to draft and validate MLCommons Croissant metadata records.

Tags

dev tools

Description

Croissant Toolkit is an intelligent suite of tools designed to work seamlessly with the MLCommons Croissant metadata format for machine learning datasets. Powered by the incredible capabilities of Google Gemini 3, this toolkit simplifies, validates, and enhances ML dataset preparation, discovery, and exploration.
codata/croissant-toolkit★ 92 forksPythonCreated Mar 14, 2026 · last push Apr 26, 2026

README excerpt

# 🥐 Croissant Toolkit


[![Event](https://img.shields.io/badge/Event-Gemini_3_Hackathon_Paris-4285F4?style=for-the-badge&logo=google)](https://ai.google.dev/)
[![Status](https://img.shields.io/badge/Status-Hackathon_Prototype-FFC107?style=for-the-badge)]()
[![License](https://img.shields.io/badge/License-MIT-4CAF50?style=for-the-badge)](LICENSE)

> **Built with ❤️ at the Gemini 3 Hackathon in Paris** 🇫🇷

**Croissant Toolkit** is an intelligent suite of tools designed to work seamlessly with the [MLCommons Croissant](https://mlcommons.org/croissant) metadata format for machine learning datasets. Powered by the incredible capabilities of **Google Gemini 3**, this toolkit simplifies, validates, and enhances ML dataset preparation, discovery, and exploration.

---

## 🌟 Key Features

* 🧙 **Wizard Data Integrator**: A single command to rule them all. Orchestrates transcription, translation, and NLP to generate fully-enriched Croissant metadata.
* 🤖 **Intelligent Metadata Generation**: Automatically generate and enrich Croissant `.jsonld` metadata from raw dataset files using Gemini 3's advanced multimodal reasoning.
* 🥐 **Croissant Expert Logic**: Deep integration with the MLCommons Croissant specification for 100% compliant JSON-LD serialization.
* 🌐 **Automated Browser Navigation**: Seamlessly launch Google Chrome and perform Google searches directly from the toolkit.
* 🎥 **YouTube Video Discovery**: Search and extract structured video data (titles, descriptions, URLs).
* 📝 **Au

Team