Heartwake

Listen to your heart and wake up naturally, every day.

Comment

Inspiration

Waking up at the same time every day should make mornings consistent, yet some days feel refreshing while others feel painfully heavy. I wondered why. After diving into sleep research, I learned how waking up during the wrong sleep phase can trigger grogginess, poor mental clarity, and low energy.

That insight sparked HeartWake, an intelligent alarm system designed to help people wake up at the right moment, not just the scheduled one. Our goal was to combine health signals with smart timing to create smoother, more natural mornings.

What it does

HeartWake is a smart alarm app that analyzes real-time user signals, such as motion, audio cues, and optional heart-rate data from wearables, to estimate sleep phases.

Instead of waking you at a rigid time, HeartWake identifies when you are transitioning through lighter stages of sleep within a custom wake-up window. If you set the latest wake time to 8:00 AM, the app may wake you earlier (like 7:50 AM) if that's the optimal moment for minimal sleep inertia and a better morning.

Key features:

  • Heart rate smart alarm concept that detects light sleep phases
  • Uses motion & sound (and optionally heart rate) to estimate sleep cycles
  • Adaptive wake-up window for a smoother morning
  • Clean, calming UI and simple user setup

How I built it

I designed HeartWake with a hybrid approach combining software engineering, UI/UX design, and applied sleep-science research.

Technologies & Process:

  • Mobile app for the main interface
  • CoreMotion for accelerometer-based sleep detection
  • Microphone analysis for breathing and movement audio patterns
  • Optional Apple Watch integration for heart rate / HRV
  • Heuristic + lightweight ML model to classify sleep states
  • Custom UI representing the heart-alarm brand identity, including our HeartWake logo

I built iteratively: start with the core logic, then add sensor processing, then integrate the UX flow and smart alarm behavior.

Challenges I ran into

  • Sleep-stage detection without wearables: Many people don’t sleep with their watch, so I had to design a fallback using motion & sound alone.
  • Real-time inference with limited battery use: Constant sensing can drain a phone quickly; optimizing that was tricky.
  • Balancing accuracy and user comfort: I needed a model simple enough for phones to run locally yet accurate enough to be meaningful.
  • iOS background execution restrictions: Implementing an alarm that monitors activity overnight required creative architecture.
  • Designing a brand identity that feels calming and intentional rather than clinical.

Accomplishments that I'm proud of

  • Creating a functional smart alarm prototype in a short time frame
  • Building a sleep-phase estimator that works without requiring a wearable
  • Designing a clean and memorable logo and UI that aligns with the app’s mission
  • Turning complex sleep-science concepts into something simple and usable
  • Proving that an early wake-up can still feel better than sleeping longer

What I learned

  • How sleep cycles influence morning energy and mood
  • The importance of thoughtful sensor fusion (motion + sound + HRV)
  • UX design principles for wellness apps
  • How to work around mobile OS limitations for background tasks
  • How small timing changes can dramatically improve the wake-up experience

What's next for Heartwake

I'm excited to develop HeartWake further:

  • Deeper machine learning models trained on multi-night datasets
  • Apple Watch + Android wearable support
  • Personalized sleep insights based on user patterns
  • Smart bedtime recommendations
  • “Energy score” predictions for next-day performance
  • Integration with HomeKit or smart lights for sunrise-style wake-ups

My long-term vision: "A fully adaptive sleep assistant that helps people wake up healthier, happier, and more naturally."

Built With

Share this project:

Updates