# Guess Your Face

> Real-time AI facial detection, 1:1 comparison, and face token analysis.
> Powered by Face++. Privacy-first, zero data retention.

Guess Your Face is a free, open-source web application that uses Face++
Cognitive Services to analyze photos in real time. All uploaded images are
processed in-memory and never stored. No login, no signup, no analytics on
user photos.

## Tools

- [Detect](https://www.guess-your-face.web.id/detect): Multi-face detection with 7 emotions (Anger, Disgust, Fear, Happiness, Neutral, Sadness, Surprise), age estimation, gender, smile, beauty score, 3D headpose, and face quality.
- [Compare](https://www.guess-your-face.web.id/compare): 1:1 face matching with confidence score and Face++ false-positive thresholds (1e-3, 1e-4, 1e-5).
- [Analyze](https://www.guess-your-face.web.id/analyze): Deep attribute inspection for up to 5 face tokens (mouth status / mask, eye status / glasses, beauty, age, emotion).

## Key facts

- Powered by Face++ Cognitive Services (US region, API v3)
- Tech stack: Next.js 16, React 19, TypeScript, Tailwind CSS v4
- Languages: Indonesian (default), English
- Zero data retention: images processed in-memory and never stored
- Max image size: 2 MB
- Supported formats: JPG, PNG
- Open source: https://github.com/REY-STTP/Guess-Your-Face

## FAQ

### Is Guess Your Face free?
Yes, Guess Your Face is free to use. There is no signup, no subscription, and no analytics on user photos. All tools run in your browser against our serverless API.

### Are my photos stored?
No. All uploaded images are processed in-memory and immediately discarded after the Face++ API responds. Nothing is written to disk, server, or database. The privacy-first architecture is verified in the open-source code.

### What is a face token?
A face_token is a unique identifier returned by Face++ for each detected face. You can copy it from Detect results and paste it into the Analyze tool to inspect deep attributes (mask, glasses, eye status, beauty) without re-uploading the original image. Face tokens are session-only and expire automatically.

### What is the difference between Detect, Compare, and Analyze?
Detect finds faces in a photo and reads 7 emotions, age, gender, smile, beauty, headpose, and face quality. Compare verifies whether two portrait photos show the same person (1:1 matching) with a confidence score. Analyze inspects previously detected faces (via face_token) for deeper attributes like mask wearing, eye status, and glasses — up to 5 tokens per request.

### How accurate is the emotion detection?
Emotion scores are produced by Face++ deep learning models trained on large datasets. Confidence varies by image quality, lighting, and face angle, but typical accuracy for the 7 emotion classes (Anger, Disgust, Fear, Happiness, Neutral, Sadness, Surprise) is high for clear frontal faces.

### Do I need to create an account?
No. Guess Your Face has no login, no signup, and no user accounts. Upload, analyze, and leave.

### Which languages are supported?
Indonesian (default) and English. Switch via the language toggle in the header. The choice is stored in your browser's localStorage.

### Which image formats and sizes are supported?
JPG and PNG only, with a maximum file size of 2 MB. Photos are cropped to 1:1 aspect ratio via an interactive canvas cropper before being sent to Face++.

### Why does the API sometimes fail?
Common reasons: image too large (>2 MB), wrong format, no face detected in the photo, Face++ quota exhausted, or rate limit hit. The UI displays a localized error message in each case.

### Is Guess Your Face open source?
Yes. The full source code is available at https://github.com/REY-STTP/Guess-Your-Face under the MIT License. You can audit the privacy claims, file issues, or fork it.

## Optional

For a longer Markdown version of this documentation, see
[llms-full.txt](https://www.guess-your-face.web.id/llms-full.txt).