JinnByteCase study

Facebloom

An AI-powered beauty and facial intelligence app that turns a single photo into personal insights.

Industry
Consumer AI and BeautyTech
Year
2023
Platforms
iOS, Android and web
Engagement
Full-stack development, product design and QA
Facebloom home screen with a face scan area and Let's Begin button
Ratings screen with overall, potential and feature scores
Overview

Turning facial intelligence into a personal, engaging consumer experience.

Facebloom uses computer vision and AI to turn a photo into facial insights, beauty guidance and social discovery. We built it across mobile and web, pairing facial analysis with highly shareable features that turned curiosity into repeat use and organic growth.

50,000+Users across the mobile ecosystem
$1M+Revenue generated
Top 20Lifestyle category on the App Store
Facebloom case studyOverview
The industry problem

Beauty advice is everywhere. Personal beauty intelligence isn't.

The same tip reaches thousands of people with completely different faces. And for image-based apps, novelty fades fast: an AI scan is interesting once, but a product needs more than that.

Problem one

Personalisation

Generic content ignores each face's proportions and features.

Problem two

Engagement

To become a business, the experience has to give people reasons to:

Explore furtherCome backCompareShareUnlock moreInvite others
Facebloom case studyProblem
The challenge

Computer vision for everyday users, without any of the complexity.

One photo in, understandable results out. Behind that single tap, the platform has a lot to do.

1Detect the face
2Map landmarks
3Analyse features
4Process results
5Translate into insight
6Personalise
7Make it shareable

It can't feel like raw data

Results had to feel personal, visual and instantly understandable.

It has to last beyond one scan

Enough variety to keep people engaged and support a subscription business.

Facebloom case studyChallenge
Our approach

Curiosity first, then personalisation and social play.

Facial analysis raises one instant question: "What does the system see when it looks at me?" That became the entry point.

Rather than building unrelated features, we used one intelligence layer to power several experiences, each with its own reason to open the app.

One facial intelligence engineSame scan, four uses
Understand myself

Personal insights

Face shape, symmetry and features, explained.

Improve something

Beauty guidance

Advice based on your own features.

Just curious

Celebrity-style match

Which famous faces look most like yours?

Compare socially

FaceMatch with friends

How alike are the two of you?

Facebloom case studyApproach
Solution architecture

One facial intelligence engine, many consumer experiences.

Every feature runs on the same facial-processing pipeline, so new experiences build on what is already there.

Layer 1

Image input

Take or upload a photo. Kept light, because everything depends on it.

Camera and gallery
Layer 2

Detection and mapping

Finds the face and the landmarks needed for analysis.

MediaPipe
Layer 3

AI analysis

Processes facial characteristics for insights and comparisons.

TensorFlow
Layer 4

Personalisation

Turns technical signals into feedback people understand.

Insights and guidance
Layer 5

Social and discovery

Powers celebrity matching and friend comparisons.

FaceMatch
Layer 6

Subscription

Premium access wraps deeper experiences.

Monetisation
Facebloom case studyArchitecture
A short path to first value

From opening the app to your results in three steps.

The first scan is the acquisition moment, so the path to it had to be as short as possible.

StartOne clear call to action on the home screen
Tell us a littleName, age, gender and hair type
Take a front selfieRetake or continue in one tap
Home screen with Let's Begin
Profile details form
Front selfie capture screen
Facebloom case studyOnboarding
Making computer vision understandable

Sophisticated underneath, simple on screen.

Nobody wants coordinates or model outputs. They want to know what those signals mean for them, so the technical layer stays hidden.

Canthal tiltPositive
Face shapeDiamond
Eye shapeAlmond
Analysis screen with canthal tilt, face shape and eye shape
Ratings with descriptions for jawline, cheekbones, nose and lips
Raw values are illustrative
Facebloom case studyInsights
Overall score result
Feature ratings for overall, potential, jawline, cheekbones, nose and lips
Results people understand at a glance

Visual scores for structure, symmetry and features.

An overall score leads into ratings for individual features and a "potential" score, which gives users something to work towards and a reason to scan again.

Overall and potentialWhere you are, and where you could be
Feature by featureJawline, cheekbones, nose and lips
HistoryCompare results over time
Facebloom case studyScores
Personalised beauty guidance

From analysis to something people can act on.

A scan is more useful when it leads somewhere. Facebloom uses facial characteristics as context for styling and self-improvement guidance, instead of giving everyone the same content.

"What does my face look like?"
"What could I try, based on my own features?"

That shift moves the product from novelty toward lasting value.

Recommendations such as face-shape and hair and beard styling
Facebloom case studyGuidance
FaceMatch result showing a facial similarity score and feature breakdown
Entertainment and social play

A story people want to tell someone else.

Consumer AI grows when its output is worth sharing. The same facial technology powers two playful experiences on top of the core analysis.

Celebrity-style matchingFind the reference faces most similar to yours. Immediate, personal and naturally shareable.
FaceMatch with friendsCompare two faces for a similarity score and the features you share most.
Facebloom case studyFaceMatch
Designing social growth into the product

FaceMatch gives every user a reason to bring a friend.

Comparing yourself with another person changes the loop entirely. Growth no longer depends only on ads or app-store discovery; the feature itself introduces Facebloom to new people.

Built-in invitationsRefer a Friend and referral codes on the home screen
Lower acquisition costProduct-led sharing reduces reliance on paid growth
1A user scansGets their own results
2Runs a FaceMatchNeeds a friend's face
3The friend joinsCurious about their own score
4They share it onAnd the loop repeats
Product-led
growthno ad spend required
Facebloom case studyGrowth loop
Designing for repeat engagement

One AI result is novelty. Many reasons to return make a product.

Acquisition

The first scan

The moment that brings people in.

Explore

Personal insights

Reasons to dig into the analysis.

Act

Beauty guidance

Reasons to come back for advice.

Grow

Self-improvement

Value beyond the scan itself.

Play

Matching and FaceMatch

Entertainment worth sharing.

Deepen

Premium access

A path for users who want more.

Facebloom case studyEngagement
From AI feature to consumer business

Making the technology work was half the job. Making it commercial was the other.

Novelty earns the first tap, not retention or revenue. Facebloom combined four ingredients.

Useful personalisation

Insights and guidance about the individual.

Entertainment

Experiences people enjoy for their own sake.

Social mechanics

Reasons to involve other people.

Premium value

Clear ways to go deeper, commercially.

=
The resultA consumer product with real market traction
Facebloom case studyBusiness model
The business impact

AI as the foundation of a growth product.

50,000+users reached across the mobile ecosystem
$1M+in revenue from engagement plus monetisation
Top 20in the App Store Lifestyle category

Product-led social engagement

Matching and FaceMatch added a social layer to the core utility.

Personalisation at scale

Individual experiences without writing guidance for every user.

Subscription-led model

Premium features monetise deeper engagement.

Facebloom case studyImpact
The outcome

From a facial scan to a scalable consumer AI experience.

CaptureTake or upload a photo
UnderstandClear insights and scores
Personalise and exploreGuidance, matches and comparisons
ContinuePremium and repeat engagement

AI personalisation, curiosity and social play in one commercially proven app: 50,000+ users and over $1M in revenue.

Facebloom case studyOutcome
Solution ecosystem

Five layers of one consumer AI product.

Facial intelligence

  • Face detection
  • Landmark mapping
  • Structural analysis
  • Symmetry insights
  • Visual processing

Personalisation

  • Appearance insights
  • Beauty guidance
  • Self-improvement content

Consumer engagement

  • Celebrity-style matching
  • Results exploration
  • Lifestyle experience

Social experience

  • FaceMatch with friends
  • Facial similarity
  • Match scoring
  • Shareable results

Commercial layer

  • Premium access
  • Subscriptions
  • Consumer monetisation
Facebloom case studyEcosystem
Technology behind the solution

Mobile-first computer vision, fast results.

MediaPipe handles face detection and landmarks, TensorFlow powers the analysis, and the Node.js backend connects results with accounts, app logic and the personalised experience.

MobileFlutteriOS and Android from one codebase
WebReactWeb experience
BackendNode.jsAccounts, results and app logic
DataMySQLUsers, scans and subscriptions
ML and visionTensorFlow and MediaPipeDetection, landmarks and analysis
Facebloom case studyTechnology
JinnByte
Facebloom

Designed and built by JinnByte.

Product design, full-stack development and QA for a consumer AI app that reached 50,000+ users and $1M+ in revenue.

Built byjinnbyte.ai
Thank you
1 / 19