Spot Me Sis (SMS)

Your AI-powered gym bestie for real-time form feedback, emotional hype, and Gen Z vibes.

Comment

Inspiration

We built Spot Me Sis because starting a fitness journey can be overwhelming, especially for women who feel uncertain about form, gym culture, or where to even begin.

We kept asking: What if working out felt like Face Timing your bestie… but she had AI superpowers?

From our own struggles with confidence, form, and staying consistent, we created an app that combines real-time form feedback, emotional support, and a touch of Gen Z fun, all powered by computer vision and machine learning.

What it does

Spot Me Sis is a mobile fitness companion that watches your form through your phone’s camera and gives live feedback, not just technically, but emotionally.

  1. Real-Time Form Feedback: Using your front-facing camera, the app detects key joints and angles to evaluate your squat or lunge form and gently coaches you as you go.
  2. Choose Your Coach: Pick from three personalities: Soft Girl, Tough Love, or Technical, each delivering form corrections in their own vibe.
  3. Track Your Glow-Up: Save your pose snapshots to a progress gallery, with time-stamped visuals of your improvements.
  4. Daily Lock-In: Set your intention, get a reminder, and start your movement streak because consistency is key.
  5. SMS-style Messages: The app speaks in your language (literally) with warm, emoji-laced encouragement or sharp one-liners (depending on your chosen coach).

How we built it

Frontend: React 18 + TypeScript + Tailwind CSS Camera & CV: MediaPipe + TensorFlow.js for real-time body tracking and angle detection Real-Time Feedback: Custom scoring logic for form quality, overlaid with live coaching messages Coach Personalities: Soft Girl, Tough Love , and Technical , powered by pre-written tone-specific feedback Tracking & UX:

  • Pose skeleton overlay
  • Score-based feedback bubbles
  • Daily Lock-in flow
  • Reminder + SMS simulator
  • Progress gallery with pose snapshots

Challenges we ran into

Smoothing out MediaPipe detection in varied lighting and room conditions Getting accurate joint angle calculations with noisy camera data Designing AI feedback that feels like a friend, not like a robot Creating a mobile layout that’s fun, functional, and not overwhelming

Accomplishments that we're proud of

Building a working real-time CV pipeline entirely in the browser Creating a UI that feels personal and emotionally supportive, not clinical Designing a feedback engine that adapts tone while staying informative Making fitness tech that genuinely feels like a Gen Z friend

What we learned

How to use Media Pipe Pose and TensorFlow.js for accurate, real-time pose detection. The value of pairing technical feedback with emotional encouragement. The importance of design, making fitness feel affirming, not robotic. How to structure and animate a mobile-first UI using React, Tailwind CSS, and Framer Motion.

What's next for SMS - Spot Me Sis

  • Add more exercise types and help with form for gym machines
  • Different girl themes (tom boy, girls girl, etc) and also boy themes
  • Launch on Expo for broader access
  • Have customizable avatars & coach personalities
  • Form detection on videos for async analysis
  • Collabs with women trainers

Spot Me Sis is going to be more than a workout tool, it's a future hype woman, form coach, and fit friend all in one. We can't wait to take her further!!

Built With

Share this project:

Updates