Hey devs π
I recently built a small side project called Face Shape Detector β a free web tool that analyzes a face photo and tells you the face shape (oval, round, square, heart, etc.) using AI.
I want to share why I built it, how it works, and a few lessons I learned along the way.
Why I Built This
I noticed that many people struggle with things like:
- Choosing the right hairstyle or glasses
- Understanding face proportions
- Finding objective feedback about their appearance
Most existing tools are either:
- Locked behind paywalls
- Mobile-only apps
- Or overly complicated
So I wanted to build something that is:
- Free
- No signup
- Works instantly in the browser
What the Tool Does
The Face Shape Detector lets users:
- Upload a face photo
- Get an instant analysis of their face shape
- See a simple explanation of the result
No data is stored permanently, and the experience is designed to be as lightweight as possible.
π Live demo:
https://www.yourface.online/
How It Works (High Level)
I tried to keep the architecture simple and scalable:
Frontend: Next.js + Tailwind CSS
Backend: API routes for face analysis
AI: Face landmark detection + rule-based classification
Deployment: Vercel
The AI part focuses on extracting facial landmarks (jawline, cheekbones, forehead width), then classifying the face shape based on geometric ratios rather than black-box predictions.
This made the results:
- More explainable
- Easier to tune
- Faster to iterate on
Challenges I Ran Into
A few things were harder than expected:
1. Image Abuse & Bots
Open tools attract spam very quickly. I had to add rate limiting and verification to prevent mass uploads.
2. Perceived Accuracy
Even when the math is correct, users may feel the result is wrong. UX explanations matter as much as the model itself.
3. Speed vs Accuracy
More complex models = slower results. I optimized for fast feedback first, accuracy second.
What I Learned
- Simple tools with clear value outperform complex ones
- UX clarity matters more than model complexity
- SEO + tool-based content is a powerful combo
- Shipping fast > waiting for βperfectβ
This project also helped me validate whether AI-powered micro tools are worth building as indie projects β and so far, the answer looks like yes.
Whatβs Next
Possible next steps:
- Hairstyle or glasses recommendations
- API access for developers
- More explainable visual overlays
If youβre interested, feel free to try it and share feedback π
https://www.yourface.online/

Top comments (0)