✨ Project Story – Sentira: The Emotionally Intelligent Ride Companion
🔥 Inspiration
During urban commutes, we often feel overwhelmed, stressed, or simply disconnected. Most micro-mobility apps focus only on efficiency — getting from Point A to B. But what if your commute understood how you felt and adapted your journey to support your mental and emotional well-being?
That’s where Sentira was born — a personal AI-powered ride companion that detects how you're feeling and turns your route into an emotionally attuned experience.
💡 What it does
Sentira uses AI to detect your mood through facial expression and voice tone analysis. It then:
- Suggests personalized ride routes (calming, scenic, energetic, or exploratory)
- Plays GPS-triggered, mood-matched audio stories
- Allows community storytelling: drop or discover stories tied to places
- Tracks your emotional shift from start to finish
⚙️ How we built it
- Next.js (TypeScript) – Full-stack framework for building the app
- Tailwind CSS + ShadCN UI – Clean, responsive, accessible UI
- Genkit + Google Gemini API – Used for generating AI-powered location-based stories
- Zod – Validates AI outputs for safe rendering
- Mediapipe – Facial expression analysis for mood detection
- Whisper API (Mocked) – Voice tone analysis (optional)
- Mapbox – Renders personalized routes on the map
- Firebase – Stores user-submitted stories and session logs
🧠 Challenges we ran into
- Balancing mood detection accuracy vs user privacy
- Ensuring real-time GPS-based storytelling without battery drain
- Making Genkit outputs reliably structured and readable with Zod
- Integrating multiple APIs without affecting app responsiveness
🚀 What's next
- Integrate Spotify’s mood-based music suggestions
- Enable live group ride experiences with shared mood
- Add user profile with mood history visualizations
- Expand local content partnerships for deeper storytelling
Log in or sign up for Devpost to join the conversation.