Private & local
All detection runs in your browser using WebAssembly and TensorFlow.js. No files are uploaded to any server — your data stays on your device, always.
File Type Checker
Detect file types instantly using AI. Drag and drop files to identify their true content type with Magika deep learning. 100% private, runs in your browser.
File Type Checker
Drop any file to identify its true type with AI. All detection runs locally — your files never leave your device.
Drag files here or click to browse
Any file type. Detection happens entirely in your browser — no upload, no server.
How to use
- Drag and drop files onto the upload area, or click to browse.
- The Magika AI model (~4 MB) loads once in your browser for accurate detection.
- Results appear below — type, MIME, and confidence score — all processed locally.
Detected Types
0 files File name Type MIME type Confidence
Features
What makes the File Type Checker reliable and easy to use.
AI-powered accuracy
Magika uses a deep learning model trained on over 100 million files across 200+ content types, achieving ~99% accuracy — far more reliable than extension-based or magic-byte methods.
Supports any file type
Drop any file — images, documents, archives, executables, scripts, and more. Magika recognizes a wide range of formats regardless of extension or metadata.
Frequently asked questions
Common questions about file type detection and how this tool works.
How does file type detection work?
This tool uses Magika, a deep learning model developed by Google that analyzes file content at the byte level. Unlike traditional methods that rely on file extensions (which can be wrong) or magic bytes (which only work for a few formats), Magika reads the raw bytes of a file and uses a neural network to predict the content type with high accuracy.
Is my file data sent to any server?
No. Everything runs locally in your browser. The Magika model is downloaded once (about 4–5 MB) and all detection happens using your device's CPU — your files never leave your computer. You can even disconnect from the internet after the model loads and the tool will continue to work.
What types of files can it detect?
Magika recognizes 200+ content types including images (JPEG, PNG, GIF, WebP, AVIF, BMP, etc.), documents (PDF, DOCX, EPUB, etc.), archives (ZIP, GZIP, TAR, etc.), executables (ELF, PE, Mach-O, APK, DEX), programming languages (Python, JavaScript, C, Go, Rust, etc.), audio/video formats, fonts, and many more.
What do the confidence scores mean?
The confidence score (0–100%) indicates how certain the AI model is about its prediction. A score of 99%+ means the model is very confident. Lower scores may indicate unusual or ambiguous file content. The score reflects the model's internal certainty — not a measure of file quality or integrity.
Why does the first detection take longer?
The first detection triggers a one-time download of the Magika deep learning model (~4–5 MB) and its TensorFlow.js runtime. Once loaded, subsequent detections are near-instant — typically completing in milliseconds. The model stays cached in your browser and won't need to be downloaded again on future visits.