Face shape detection for your product
One REST endpoint. Send a photo, get confidence scores for all six face shapes back in seconds. Built for eyewear shops, salon software, and beauty apps.
Live on RapidAPI · JPEG or PNG · No images stored
curl -X POST https://api.detect-face-shape.com/detect \
-H "X-RapidAPI-Key: YOUR_KEY" \
-F "image=@portrait.jpg"- oval
- 0.61
- round
- 0.14
- square
- 0.09
- oblong
- 0.08
- heart
- 0.05
- diamond
- 0.03
How it works
Three steps from photo to product feature.
Send a photo
POST an image to /detect. JPEG or PNG up to 5MB, one face per shot.
The model scores it
A vision transformer trained on labeled faces rates how well each of the six shapes fits.
You get JSON back
Six confidence scores that sum to one, ready to drive recommendations in your app.
What you can build
The endpoint stays the same. What you build on top is up to you.
Read the API docs-
Eyewear stores
Recommend frames that suit each visitor's face shape, right on the product page.
-
Salon and hairstyle apps
Match haircuts to face shapes before the client sits down in the chair.
-
Beauty and analysis tools
Add face shape results to quizzes, virtual try-ons, and personal styling reports.
Choose the right plan
Both plans run on RapidAPI. Pick the scale that matches your application.
Pro Plan
- Pay as you go, no monthly fee
- Rate limit: 10 requests per second
Ultra Plan
Most popular- 6,000 requests per month included
- $0.04 per additional request
- Rate limit: 10 requests per second
Frequently asked questions
Your questions about the Detect Face Shape API, answered.
What are the image requirements?
Images should be JPEG or PNG, at most 5MB, with a minimum resolution of 200x200 pixels for accurate face shape detection.
How fast does the API respond?
Response times vary with image size and server load, but you can typically expect a response within a few seconds.
Are there any rate limits?
Yes. Both plans are limited to 10 requests per second to keep performance fair for all users.
How accurate is the detection?
The API uses a vision transformer model that delivers high accuracy. Image quality and how the face is positioned still influence the result.
Can it handle images with multiple faces?
Not yet. The API currently analyzes one face per image. Multi-face support is on the roadmap.
What about privacy and data protection?
Uploaded images are processed for face shape detection only. They are not stored and not used for anything else.
Still have a question?
If the FAQ did not cover it, the full documentation on RapidAPI has request samples, schemas, and support.