I built Palette by Momo (https://www.palettebymomo.com) for fun, to democratize seasonal color analysis. Traditional color consultations cost $100-300 and require in-person appointments, but I wanted something instant and accessible that anyone could use from their phone.
Seasonal color analysis helps people find the most flattering colors for their skin tone, hair, and eyes. It's incredibly popular on TikTok and Instagram, but professional consultations are geographically limited and costs $100 to $300. I tried a couple of other apps and my attempts often fail because it's genuinely difficult to assess your own coloring objectively.
I integrated the SkinToneClassifier library (https://github.com/ChenglongMa/SkinToneClassifier), which uses face detection, skin segmentation, and k-means clustering to determine skin tone categories. The library is actually quite elegant - it extracts dominant colors from detected facial areas and maps them to established color palettes used in professional analysis. I used FastAPI for the skintone analysis backend and really liked the developer experience.
The FastAPI endpoint processes the uploaded image, runs it through the skin tone classifier, and returns a comprehensive color palette tailored to the user's analysis.
How It Works:
Upload a photo → User submits a clear photo of their face
AI analysis → The system analyzes skin tone, hair color, and eye color
Personalized results → Get a comprehensive custom palette.
The entire process takes about 2 minutes and users love having instant access to personalized color analysis that would otherwise require expensive appointments. It's a perfect example of how AI can make specialized knowledge accessible to everyone.
Seasonal color analysis helps people find the most flattering colors for their skin tone, hair, and eyes. It's incredibly popular on TikTok and Instagram, but professional consultations are geographically limited and costs $100 to $300. I tried a couple of other apps and my attempts often fail because it's genuinely difficult to assess your own coloring objectively.
I integrated the SkinToneClassifier library (https://github.com/ChenglongMa/SkinToneClassifier), which uses face detection, skin segmentation, and k-means clustering to determine skin tone categories. The library is actually quite elegant - it extracts dominant colors from detected facial areas and maps them to established color palettes used in professional analysis. I used FastAPI for the skintone analysis backend and really liked the developer experience.
The FastAPI endpoint processes the uploaded image, runs it through the skin tone classifier, and returns a comprehensive color palette tailored to the user's analysis.
How It Works: Upload a photo → User submits a clear photo of their face AI analysis → The system analyzes skin tone, hair color, and eye color Personalized results → Get a comprehensive custom palette.
The entire process takes about 2 minutes and users love having instant access to personalized color analysis that would otherwise require expensive appointments. It's a perfect example of how AI can make specialized knowledge accessible to everyone.