GitHub - google/magika: Fast and accurate AI powered file content types detection
6 hours ago
- #file-detection
- #open-source
- #deep-learning
- Magika is an AI-powered file type detection tool using deep learning, optimized for accuracy with ~99% average accuracy across 200+ content types.
- It processes files quickly, with inference times of about 5ms per file on a single CPU, and handles large volumes, including hundreds of billions weekly for Google services like Gmail and Drive.
- Magika is available as a command line tool, Python API, and bindings for Rust, JavaScript/TypeScript, and Go, with features like recursive scanning and customizable confidence modes.
- It includes a web demo, open source components under Apache 2.0 license, and is used for safety enhancements in Google products and integrated with platforms like VirusTotal.