Ai Powered Maternal Health Smart

Supporting Mothers, Saving Lives, Anytime, Anywhere

Comment

AMA Pregnancy Health Assistant – Documentation

Empowering Safe and Informed Pregnancy Journeys with AI-Driven, Offline-Ready Healthcare Technology

🌟 Inspiration

Across Africa, especially in underserved rural communities, maternal mortality remains one of the most pressing health challenges. Many mothers face life-threatening situations due to lack of access to timely medical advice, healthcare facilities, and emergency assistance.

We've seen countless heartbreaking stories where women lost their lives from entirely preventable complications—simply because they didn’t know the warning signs or couldn’t access help in time.

We asked ourselves:
What if every mother—regardless of where she lives or her level of literacy—could access personalized, life-saving maternal health guidance, even without an internet connection?

That’s how AMA (AI-powered Maternal Health Assistant) was born—a mission-driven platform designed to save lives by leveraging the power of offline-first mobile technology and intelligent web systems.

💡 What AMA Does

AMA is a cross-platform maternal health solution consisting of:

  • A native mobile app (built with Flutter)
  • A responsive web application (built with React)

Together, they deliver a consistent, reliable, and intelligent healthcare experience to both pregnant women and healthcare providers.

🔑 Key Capabilities:

🤖 Voice-Assisted AI

  • Natural conversation interface for maternal health Q&A
  • Speech-to-text and text-to-speech in Amharic and English
  • Context-aware responses based on symptoms, pregnancy stage, and location
  • Works offline on mobile

🧠 Offline-First Health Guidance

  • WHO-approved maternal health knowledge base
  • Smart alerts for danger signs and emergency escalation
  • Custom reminders for checkups, supplements, and vaccinations
  • Secure local storage with offline sync

📊 Health Passport and Monitoring

  • Full medical record storage and retrieval
  • QR code for easy data sharing with healthcare providers

🚨 Emergency & Danger Detection

  • AI-powered risk predictor to trigger referrals or alerts
  • SMS fallback for critical conditions
  • 24/7 danger detection with custom alert rules

📈 Pregnancy Journey Tracking

  • Weekly check-ins and due date calculations
  • Personalized health recommendations
  • User-controlled logs of symptoms and mood

🔐 Secure User & Data Management

  • Supabase-backed authentication (web)
  • Secure local user sessions (mobile)

🛠️ How We Built AMA

The platform was built to support:

  • Low-connectivity environments
  • Multilingual voice interaction
  • Real-time data synchronization

Accessible on both smartphones and desktop browsers.

🔧 Cross-Platform Development Stack

Layer Technologies
Frontend (Mobile) Flutter · Dart · SQLite · AddisAI · Vosk STT/TTS
Frontend (Web) React 18 · TypeScript · TailwindCSS · Shadcn/UI · React Router
Backend (Both) NestJS · Node.js · Supabase (DB/Auth/Storage) · PostgreSQL
Voice AI AddisAI for Amharic voice · Web Speech API
Data Sync Offline-first mobile DB sync
Security & Privacy Local encryption · Supabase RLS · HIPAA-inspired data structure

⚙️ Key Features in Action

🎙️ Voice Technology

  • Mobile: AddisAI with Vosk engine for offline speech-to-text
  • Web: Web Speech API for real-time voice interaction

🤖 Custom AI Health Assistant

  • Lightweight LLM focused on maternal health
  • Adjusts dynamically based on user profile (e.g. pregnancy week)
  • Maintains multilingual flow (Amharic & English)

📲 Health Monitoring

  • QR code generation for data sharing
  • Syncing with Supabase on web
  • Background job queuing for offline uploads on mobile

⏰ Reminders & Emergency Detection

  • Local scheduled reminders (mobile)
  • Cron-based services (web)
  • Risk scores trigger alerts and SMS (Twilio/Webhooks or native messaging)

🚧 Challenges We Faced

🔄 Offline & Online Seamlessness

  • Robust local caching & sync strategies
  • Sync conflict resolution via background jobs

🗣️ Voice & Language

  • Supporting Amharic dialects
  • Culturally adaptive phrasing in AI responses

🌍 Content Localization

  • WHO-compliant but simplified content
  • Collaboration with local health professionals

🔐 Security & Trust

  • End-to-end data encryption
  • HIPAA-inspired health data handling

🎉 What We're Proud Of

  • 🔒 Fully offline mobile functionality for rural areas
  • 🎙️ Multilingual voice assistant without internet
  • ❤️ Emergency detection & alerts for high-risk cases
  • 📱 Unified Health Passport via QR code
  • 🌍 Inclusive design for low literacy & regional coverage

🧠 What We Learned

Technical Insights

  • Offline-first apps must fail gracefully and sync smartly
  • Voice UX is highly contextual and culture-sensitive
  • Consistency across devices builds user trust

Human-Centered Design

  • Design must cater to low digital literacy
  • UI must be empathetic and intuitive
  • Direct feedback is vital for solving real user needs

🔮 What’s Next

🛣️ AMA Integration Roadmap

  • Bluetooth Diagnostics
    Connect devices (e.g., fetal monitors) to sync with web dashboards

  • More Languages
    Add Oromo, Tigrigna, and Swahili

  • Health Worker Dashboards
    Web-based analytics for maternal health trends

  • Public Health Campaigns
    Use anonymized data for outreach and intervention

  • Community Training
    Collaborate with NGOs to train health workers

👥 Project Team

Lidetu Ketema – 0972315453
Dawit Gerim – 0985392862

"AMA is more than an app—it's a movement to safeguard motherhood through the power of inclusive AI and compassionate technology."

Built With

Share this project:

Updates

Submission history