Detect. Calm. Heal. — Real-time stress detection and AI-powered relief, right on your wrist.
Alleaf is a wellness app that listens to your body so your mind doesn't have to fight alone.
Using wearable sensors that monitor heart rate and galvanic skin response (GSR), Alleaf detects when your stress levels spike — and responds instantly. It triggers haptic bilateral stimulation (alternating vibrations inspired by EMDR therapy) to ground you in the moment, while an AI therapy agent guides you through personalized mindfulness exercises tailored to what your body is telling it.
No waiting rooms. No scheduling. Just calm, on demand.
- 🔬 Real-Time Stress Detection — Continuous monitoring of heart rate and GSR signals via wearable sensors, processed through a trained scikit-learn classification model.
- 📳 Haptic Bilateral Stimulation — Alternating vibration patterns delivered to calm the nervous system, grounded in the neuroscience behind EMDR therapy.
- 🤖 AI Therapy Agent — A LangChain-powered conversational agent that conducts supportive check-ins and adapts its approach to your current stress profile.
- 🧘 Agentic Mindfulness Exercises — Dynamic, context-aware exercises that evolve based on your biometric state — not a one-size-fits-all playlist.
- 📊 Stress Trend Insights — Pandas-powered analytics that surface patterns in your stress over time so you can understand your triggers.
| Layer | Technology |
|---|---|
| Frontend | React / Next.js |
| Backend | Python / Flask |
| ML & Signal Processing | scikit-learn, Pandas |
| AI Agent | LangChain |
| Sensors | Wearable heart rate + GSR |
- Node.js ≥ 18
- Python ≥ 3.10
- A compatible wearable device
# Clone the repository
git clone https://github.com/Gustav-Galvao-e-Silva/alleaf.git
cd alleaf
# Install frontend dependencies
cd client
npm install
# Install backend dependencies
cd ../server
pip install -r requirements.txt# Start the Flask backend
cd server
flask run
# Start the Next.js frontend (in a separate terminal)
cd client
npm run devThen open http://localhost:3000 and pair your wearable to begin.
Wearable Sensors (HR + GSR)
│
▼
Stress Classification (scikit-learn)
│
┌────┴────┐
▼ ▼
Haptic LangChain AI Agent
Bilateral → Mindfulness Exercise
Stimulation → Therapeutic Check-in
When elevated stress is detected, Alleaf triggers bilateral haptic feedback immediately while simultaneously prompting the AI agent to initiate a calming session. The agent adapts its tone and exercises based on the severity of the detected stress signal.
Mental health support is still largely reactive — you seek help after a crisis. Alleaf flips that model. By sitting quietly in the background and responding the moment your body shows signs of distress, it brings proactive, evidence-based care into everyday life.
Bilateral stimulation has strong roots in clinical therapy (EMDR). We're making it accessible, automatic, and personal.
This project was built during Hackalytics in 2026.
| Name | Role |
|---|---|
| Deep | Frontend |
| Gustavo | ML |
| Jhoon | Hardware |
| Kalyan | Backend |
This project used the publicly available WESAD dataset: Philip Schmidt, Attila Reiss, Robert Duerichen, Claus Marberger and Kristof Van Laerhoven, "Introducing WESAD, a multimodal dataset for Wearable Stress and Affect Detection", ICMI 2018, Boulder, USA, 2018.