SENTINA

• 

  app logo based on featured functionalities

Inspiration

Our inspiration comes from a deeply personal place. Recently, a team member's grandmother, who raised them, became critically ill in the ICU. The sheer physical and emotional toll on family caregivers became painfully apparent. We realized that while many health-tech solutions focus entirely on the patient, very few address the emotional, mental, and time stress of the guardian. Sentina was born to share this weight—providing a digital safety net for Alzheimer's patients while offering profound emotional and practical support for the caregivers who love them.

We asked:

What if AI could feel like family?

Sentina was created to restore a sense of safety and connection through personalized, emotionally intelligent AI.

What it does

Sentina is a comprehensive companion and monitoring web app designed specifically for Alzheimer's/dementia patients and their guardians. It features three core pillars:

Location Tracking: Real-time monitoring to ensure the patient stays within designated safe zones.

Talk with Me (Voice Companion): An AI-driven conversational interface. Patients can talk to the app, and it responds in the cloned, familiar, and comforting voice of their loved one. This helps answer repetitive questions and soothe the patient without causing caregiver burnout.

Caregiver Statistics: An intuitive dashboard that tracks daily out-of-zone safety incidents and logs chat history. Crucially, it automatically detects and flags "repetitive

How we built it

We built the frontend using React and react-router-dom for seamless Single Page Application (SPA) navigation, styled entirely with Tailwind CSS to create a calming, ocean-wave-inspired glassmorphism UI. We integrated Firebase for secure user authentication.

The core "Talk with Me" feature relies on a powerful AI pipeline:

We use OpenAI's Whisper API for accurate speech-to-text processing.

GPT-4o serves as the empathetic brain to understand context and generate appropriate responses.

ElevenLabs API provides state-of-the-art voice cloning to synthesize the caregiver's voice in real-time.

Challenges we ran into

Orchestrating the real-time AI voice pipeline was incredibly demanding. Google Cloud, Gemini (LLM), and ElevenLabs (TTS) while maintaining low latency to ensure a natural conversation flow required significant optimization. On the frontend, managing complex React states and nested routing across our highly detailed Statistics dashboard (including dynamic calendars and daily interaction logs) was a major UI/UX hurdle. Additionally, navigating the ethical implications of voice cloning consent and planning for HIPAA-compliant data storage required careful architectural planning.

Accomplishments that we're proud of

We are incredibly proud of shifting the paradigm from "patient-only" care to "caregiver empowerment." We are specifically thrilled with our Statistics Dashboard—it doesn't just display raw data; it intelligently flags repetitive questions to help caregivers track disease progression over time. We are also deeply proud of the exceptionally smooth, calming UI. We successfully avoided a sterile, "clinical" look, creating an app that feels like a warm, supportive companion.

What we learned

From a technical standpoint, we learned how to seamlessly integrate complex audio-processing and voice-synthesis APIs into a cohesive React application. More importantly, we learned the true value of Empathy-Driven Development. Every single design choice, from the soothing background animations to the specific comforting messages displayed on the caregiver's dashboard, was a lesson in how software can provide genuine emotional relief.

What's next for SENTINA

Moving forward to v1.5 and v2.0, we plan to implement clinical validation for our cognitive trend scoring via academic partnerships. We also aim to upgrade our browser-based GPS to a native mobile app framework for more accurate indoor tracking and background geofencing. Finally, we are designing a robust BAA framework with end-to-end encryption to meet stringent HIPAA safeguards for storing voice transcripts and movement logs.

Built With

• api
• css
• html
• javascript
• react
• tailwind