An internal AI facial analysis app, built by Hoop.
Facelyzer is an internal app: a mobile product that takes a single face photo and returns a reliable face-shape reading, then turns it into styling guidance. Our team built the Flutter app and the OpenCV-based analysis backend behind it in two months.
- Project Type
- Internal App
- Industry
- AI Beauty & Styling
- Timeline
- 2 Months
- Status
- Completed
In short
Facelyzer is a facial analysis application that detects facial structure and shape using computer vision and image-processing pipelines, enabling automated analysis for styling, personalization and digital recommendations. We built the Flutter mobile app and the OpenCV-based analysis backend, designed the flow through our product design process, and delivered the core build in two months.
- Flutter frontend application for iOS and Android
- OpenCV-based analysis backend
- Image-processing pipeline for feature extraction
- Processing performance optimized for mobile usage
Computer vision that has to work on any phone
A face-shape reading is only useful if it is right, and the input is the hardest part: users photograph themselves on very different device cameras, in different light, and expect a result quickly. The brief was to make that dependable inside a mobile app rather than a lab demo.
On top of detection itself, the analysis had to plug into an ordinary mobile workflow, and the raw reading had to become something a user can act on.
Accurate face detection across device cameras
Different cameras produce very different photos, and the analysis had to stay reliable across them.
Real-time image processing performance
Image analysis is demanding, and it had to feel workable on a phone rather than a workstation.
Integrating ML pipelines into mobile workflows
The analysis pipeline had to fit naturally into an app’s upload, wait and result flow.
Turning a reading into a usable result
A detected shape needed to become ranked match percentages and styling tips a user can act on.
A Flutter app on top of an OpenCV analysis backend
We developed the Flutter frontend application to handle capture, upload and results on both platforms, and implemented an OpenCV-based analysis backend to do the demanding work. Between them sits an image-processing pipeline built for feature extraction.
Processing performance was then optimized for mobile usage, so the app stays responsive while the analysis runs. The result is a system that takes an uploaded photo and returns a face-shape reading the rest of the product builds on.
From upload to face shape, screen by screen
Five screens carry the whole experience: upload a photo, wait while the analysis runs, read the result, ask the AI chat, and see what premium adds.
Step 1
Upload
Add a face image, keep the face within frame
Step 2
Processing
Analysis runs in the background; users can come back later
Step 3
Result
Face shape, match percentages and styling tips
Step 4
AI Chat
Style questions with quick suggestion chips
Step 5
Premium
Plans, free trial and full access
The feature set across four areas
The features below are the ones visible across the app's screens. The analysis core — the Flutter app, the OpenCV backend and the feature-extraction pipeline — is the part we built; everything else sits on top of it.
Analysis
- Face Image Upload With Framing Guidance
- Face Shape Detection
- Match Percentage Across 5 Shapes
- Background Processing State
Styling
- Face-Shape Styling Tips
- Hair & Beard Suggestions
- Celebrity Face Match
- Glasses Recommendations
AI Chat
- Free-Text Style Questions
- Quick Suggestion Chips
- Answers Tailored To Face Shape
Access
- Free Analysis
- Premium Tier & Get Pro Prompts
- Plan Selection & Free Trial
- Locked Tips Unlocked On Upgrade
Roadmap note: the app's home banner announces an Eye Shape Detector as coming soon.
Three pieces that make the analysis dependable
OpenCV-based analysis backend
The demanding image analysis runs in a dedicated backend built on OpenCV, keeping the mobile app light.
Feature-extraction pipeline
Every photo passes through an image-processing pipeline built to extract the facial features the shape reading depends on.
Mobile-optimized performance
Processing was tuned for mobile usage so results arrive in a way that suits how people actually use their phones.
The results
AI-Powered Facial Analysis
Functional
Image Processing Workflows
Reliable
Personalization-Driven Foundation
Extensible
Core Build Delivered
2 Months
The services behind this app
Have an AI product idea?
Tell us where you want to go. We'll show you how to get there — and then build it with you.