Inspiration
Like Pluto — once a planet, now distant and almost forgotten — our memories can fade, but they never truly disappear.
We were inspired by how Alzheimer’s patients struggle to stay connected to who they are and the people who love them. We wanted to create a space where technology could preserve what makes us human — our memories, our emotions, and our relationships.
What it does
Pluto is an AI-powered memory companion designed to support individuals experiencing memory loss, such as those with Alzheimer’s disease. It allows families and caregivers to upload photos, memories, and personal stories, which are used to build a personalized Memory Profile.
Through natural conversations with the AI chatbot, users can:
Revisit important memories through photos and stories.
Ask questions like “Who is this?” or “Where was I born?” and receive emotionally contextual answers.
Engage in self-reflective dialogue that helps anchor them in their past, present, and future self.
Access calming and positive conversations to reduce feelings of loneliness or confusion.
How we built it
Frontend:
React.js + Vite for a fast, responsive, and modular web interface.
Vercel for smooth deployment and global scalability.
Backend:
Flask API to handle chat and memory data flow.
MongoDB for storing user profiles, memory data, and caregiver inputs.
AI System:
Gemini Model (Google Generative AI) for generating emotionally intelligent responses and self-dialogue reflections.
PyTorch + scikit-learn + BERT for custom sentiment analysis, allowing Pluto to detect emotional tones (e.g., joy, sadness, confusion) and adjust its tone accordingly.
Challenges we ran into
Integrating Gemini with our Flask backend while maintaining context between user sessions.
Training a lightweight sentiment model that could run efficiently with minimal latency.
Ensuring the system handled sensitive emotional data ethically and securely.
Designing a UI that’s simple enough for elderly users while still modern and warm.
Accomplishments that we're proud of
Built a working AI chatbot that remembers and references uploaded memories.
Integrated emotional sentiment analysis to shape tone and response style dynamically.
Created a beautiful, emotionally grounded interface inspired by the theme of “past, present, and future.”
Deployed the system on Vercel + Flask, integrating multiple APIs successfully within the hackathon timeframe.
What we learned
How to integrate multiple AI systems (Gemini + BERT) for contextual and emotional awareness.
The importance of designing AI for empathy, not just accuracy.
That emotional design — tone, color, and pacing — is as crucial as technical design for accessibility and user trust.
What's next for Pluto
Voice-based interactions for users with limited mobility.
Photo recognition AI to identify people and events automatically.
Family dashboard for caregivers to monitor emotional well-being and engagement levels.
Integration with wearable devices for emotional and biometric feedback (e.g., heart rate, sleep, anxiety).
Built With
- atlasgemini-api
- bert
- flask
- html/css
- mongodb
- python
- pytorch
- react.js
- render
- scikit-learn
- tailwind
- vite

Log in or sign up for Devpost to join the conversation.