AMIE | Devpost

🧠 AMIE – Assistant for Memory, Identity, and Emotion

🔍 Overview

AMIE is a mobile application built with empathy at its core — designed to support individuals suffering from Alzheimer's disease. Leveraging personalized voice cloning, face recognition, memory prompts, and emotional intelligence, AMIE helps patients recall, reconnect, and relive their most meaningful moments.

We created AMIE not just as a tool, but as a companion — one that understands their world, brings them comfort, and helps reduce the confusion that Alzheimer’s often brings.

🧩 The Problem

Alzheimer’s patients face a cruel reality — they begin to forget not just events, but their loved ones, their purpose, and even themselves. Caregivers often struggle to communicate meaningfully, while patients cycle through confusion and frustration. Existing apps are impersonal, complicated, and offer little emotional support or personalization.

💡 Our Solution

AMIE is a personal mobile assistant that brings memory back using:

🗣️ Custom voice memory recall – AMIE uses the original voices of loved ones to narrate key memories, making recall stronger and more emotionally resonant.
🧍‍♂️ Face recognition – Instantly identifies known and unknown people using on-device AI, helping patients understand who’s around them.
🧏 One-word recall trigger – Say one word (like “wedding” or “Diwali”) and AMIE brings back all related images, voices, and emotions associated with that event.
🧠 Sentiment-driven importance detection – Prioritizes memories and reminders based on their emotional weight and how the patient reacts.
🔁 Talk-back functionality – Patients can have real conversations with AMIE, making them feel heard and engaged.
🎶 Music memory assist – Reminds users of the songs played during specific events to trigger deep emotional memory.
🔒 Event-based lock screen wallpapers – Shows photos of past events or loved ones on the lock screen and asks: “Do you remember these people?” to gently stimulate memory.

⚙️ Tech Stack

Frontend: React Native
Backend: Django Rest Framework (DRF)
Task Queue: Celery + Redis
Voice: Custom-trained voice cloning models (instead of TTS)
AI/NLP:
- Face recognition (on-device using MLKit or similar)
- Sentiment Analysis using VADER & DistilBERT
- One-word memory linking using sentence embeddings

🎯 Key Features

Voice cloning of real relatives
Memory recall based on emotional significance
Real-time known/unknown face recognition
Personalized memory triggers
Music-based memory jogger
Lock-screen event picture prompts
One-word deep recall
Sentiment-aware memory prioritization
Talk-back companion chatbot
Caregiver-friendly interface and analytics

📊 Impact & Key Metrics

🔁 Memories recalled/day
😊 Emotional uplift via sentiment tracking
🧠 Accuracy of face recognition & one-word response
👥 User engagement with memory prompts
📈 Growth in subscriptions/referrals

💰 Cost Structure

App Development: ₹1.5 lakhs
Voice Model Training & Hosting: ₹1.2 lakhs
On-device Model Tuning & Optimization: ₹1 lakh
UX Research (Therapists & Patients): ₹50k
Marketing to Health Institutions: ₹80k
Total Estimated Cost: ₹5 lakhs (~\$6K)

This covers early-stage development for around 5,000 users at launch.

📈 Business Model

💼 B2B: Partner with hospitals, elder-care homes, and Alzheimer clinics
📱 Subscription: Affordable personal monthly/yearly plan for families
🎁 Referral incentives for spreading AMIE to others
🧑‍⚕️ Therapist dashboard for emotional analysis and trend tracking

🚀 Future Scope

📲 iOS & multilingual expansion (Hindi, Marathi, Tamil, etc.)
🧠 Integration with wearable health data (heart rate, sleep patterns)
👓 Smart Glasses integration for real-time face & object recognition
🧬 Predictive memory loss tracking using AI
💌 AMIE as a companion for patients with early-stage dementia or PTSD

💖 The Story Behind AMIE

It all started when I went to visit a friend’s grandmother. She welcomed me with a kind smile and asked, “Beta, kitne log hain ghar pe?” I answered and we laughed. But just two minutes later, she asked the same question again. Then again: “Papa kya karte hain?” The loop repeated. That moment wasn’t just confusing — it was heartbreaking. She was trying to connect, to hold on, but her memory wouldn’t let her. That very evening, we started brainstorming — what if technology could fill in those gaps without making her feel lost?

That experience gave birth to AMIE — an emotional, AI-powered assistant to bring back those forgotten moments. Because no one should forget the people they love. https://youtu.be/wQDX8eWEDWs