---
id: magika
version: "1.0.3"
license: Apache-2.0
license_treatment: permissive
maintenance: active
---

# magika — A tool to determine the content type of a file with deep learning

License: permissive · Maintenance: active · Downloads: 15.2M/mo

## What it is and what it does

Magika is an AI-powered file type detection tool that uses a lightweight deep learning model to identify file content types by analyzing file bytes rather than relying on file extensions or magic numbers alone. It works as both a standalone command-line tool and a Python library, making it suitable for integration into larger systems or use as a utility.

The package wraps a custom-trained neural network model (a few MBs in size) that has been trained on approximately 100M samples across 200+ content types, covering both binary and textual formats. It provides structured output including the detected label, description, MIME type, file extensions, and a confidence score. The tool supports multiple output formats (plain text, JSON, JSONL) and can process files from disk, stdin, or byte streams in Python code.

Use it for:

- Validate uploaded file types in web applications without trusting user-supplied extensions or MIME type headers.
- Automatically categorize and organize files in bulk processing pipelines by their actual content rather than filename.
- Detect potentially malicious files disguised with incorrect extensions during security scanning workflows.
- Build content-aware indexing systems that need accurate file type information for search or filtering.
- Replace or supplement traditional file magic number detection in tools that need higher accuracy on modern formats.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Magika identifies file types using deep learning, analyzing file content to determine MIME types and file categories with high accuracy, available as both a command-line tool and Python library.

Yes. Magika is actively maintained, has low install friction, carries no security vulnerabilities, and uses a permissive license. It solves a real problem (accurate file type detection) with a well-engineered approach. Install it if you need reliable file type identification beyond simple extension or magic-byte checks; skip it only if your use case is limited to a handful of well-known formats where simpler tools suffice.

## Install

pip install magika
uv add magika
poetry add magika

## Installing magika

Before you install: Low friction install with only two runtime dependencies (click and onnxruntime). Active maintenance with recent commits and a large repository following (17950 stars). Supports Python 3.8 through 3.14, distributed as a pre-built wheel.

License in practice: Licensed under Apache-2.0 (permissive), allowing commercial and private use with minimal restrictions. Suitable for most projects without license compatibility concerns.

Quickstart:

pip install magika

from magika import Magika
m = Magika()
res = m.identify_bytes(b'function log(msg) {console.log(msg);}')
print(res.output.label)  # outputs: javascript

onnxruntime must be available; the package includes platform-specific compiled wheels for common architectures, with a pure-Python fallback for unsupported platforms.

Verify before relying:

- Whether the ~100M training samples and 200+ content types mentioned in the description are current for version 1.0.3.
- Performance characteristics (millisecond detection times) on typical hardware configurations.
- Exact list of all 200+ supported file formats and their detection accuracy breakdown.

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 15.2M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

file type detection, content type identification, deep learning file detection, MIME type detection, file format recognition, magic bytes alternative, AI-powered file analysis, file-detection, deep-learning, security

[View on SkillFed](https://skillfed.io/packages/magika) · [View on PyPI](https://pypi.org/project/magika/)