Project Title

magika — Fast and Accurate AI-Powered File Content Type Detection

Overview

Magika is a novel AI-powered file type detection tool that leverages deep learning to provide accurate detection. It employs a highly optimized model, enabling precise file identification within milliseconds, even on a single CPU. Magika has been trained on a dataset of ~100M samples across 200+ content types, achieving an average ~99% accuracy.

Key Features

Highly optimized model for fast file type detection
Trained on a large dataset of ~100M samples across 200+ content types
Achieves ~99% accuracy on test sets
Available as a command line tool, Python API, and additional bindings for Rust, JavaScript/TypeScript, and GoLang (WIP)

Use Cases

Improving Google users' safety by routing Gmail, Drive, and Safe Browsing files to proper security and content policy scanners
Integration with VirusTotal and abuse.ch for enhanced file type detection
General file type detection for various applications, including security and content policy enforcement

Advantages

Fast and accurate file type detection, even on a single CPU
Trained on a large and diverse dataset, covering both binary and textual file formats
Open-source and actively maintained by Google
Available in multiple programming languages for flexibility

Limitations / Considerations

The documentation on GitHub refers to the latest, potentially unreleased and unstable version of Magika
The latest stable release of the magika Python package is 0.6.1, which may have different features or performance characteristics
The web demo runs an older version of the model, which may yield different results compared to the latest version

Trend Micro's File Type Detection: A deep learning-based file type detection tool, but with a different focus and dataset.
Apache Tika: A content analysis toolkit that can detect file types, but without the same level of accuracy and speed as Magika.
ExifTool: A platform-independent tool to read, write, and edit meta information in a wide variety of files, but not specifically focused on AI-powered file type detection.

Basic Information

GitHub: https://github.com/google/magika
Stars: 8,833
License: Unknown
Last Commit: 2025-09-26

📊 Project Information

Project Name: magika
GitHub URL: https://github.com/google/magika
Programming Language: Python
⭐ Stars: 8,833
🍴 Forks: 451
📅 Created: 2023-08-22
🔄 Last Updated: 2025-09-26

🏷️ Project Topics

Topics: [, ", a, i, ", ,, , ", d, e, e, p, -, l, e, a, r, n, i, n, g, ", ,, , ", f, i, l, e, t, y, p, e, ", ,, , ", k, e, r, a, s, -, c, l, a, s, s, i, f, i, c, a, t, i, o, n, -, m, o, d, e, l, s, ", ,, , ", k, e, r, a, s, -, m, o, d, e, l, s, ", ,, , ", m, i, m, e, -, t, y, p, e, s, ", ,, , ", o, n, n, x, ", ]