AuraCare +

Inspiration

Millions of seniors live alone, and the people who care for them often cannot be there every moment. Medication routines, loneliness, and delayed communication can make everyday life more difficult. Most health tools feel clinical and demand that a senior learn a new interface. We wanted something that works the way seniors already prefer: just talking. AuraCare+ is a friendly presence for Elenor that listens, helps with her daily routine, and keeps her caregiver, Esther, informed when it matters.

What it does

AuraCare+ is an ambient voice companion, "Aura," built as a simulated Alexa+ experience. It gives seniors natural conversation through a simple, accessible interface.

Hands-free voice conversation: live transcription, automatic sending when the user stops speaking, and spoken replies.
Medication support: medication names, doses, times, and due-status tracking. Aura gives gentle reminders and safe guidance, and it never diagnoses or changes a prescription.
Caregiver connection: Aura helps initiate contact with Esther, records activity, and encourages escalation when something seems wrong.
Emergency-aware behaviour: for falls, chest pain, breathing trouble, or unsafe situations, Aura gives clear emergency guidance and prompts contact with emergency services and the caregiver. It never pretends to replace medical professionals.
Camera check-in controls: the user can turn the camera check-in on or off using the camera control or a voice command such as "Aura, turn on the camera."
Transparent activity feed: the interface shows listening status, captured speech, medication reminders, assistant responses, and system activity so caregivers and judges can understand what Aura is doing.
Senior-friendly design: large controls, high contrast, clear status messages, and plain language.

How it uses Alexa+ and MCP

For the Alexa+ track, we chose the simulated Alexa+ experience option: a web app built with v0 by Vercel and also used the help of google ai studio. We used it to focus on what matters most for seniors: the voice interaction, the medication workflow, the caregiver workflow, and a transparent activity feed that shows what Aura is doing.

The senior-facing capabilities are organized as tool-style actions that map directly to MCP tools:

get_medications returns today's schedule and due status.
log_checkin records a wellness check-in.
notify_caregiver alerts Esther through a connected caregiver channel.

In the current demo, these actions run through the application's API routes. The structure is designed so they can be exposed through a self-hosted Streamable HTTP MCP server as the next step, connecting AuraCare+ to Alexa+ without changing the senior-facing experience.

How we built it

Frontend: Next.js, React, and TypeScript manage the conversation, medication display, camera controls, voice input, and caregiver actions.
Voice: the browser Web Speech API handles speech recognition and text-to-speech. Spoken words appear in the text field while the user talks and are submitted when speech finishes.
Backend: Next.js API routes provide assistant responses, medication data, contextual fallbacks, caregiver actions, and safety-aware behaviour.
AI layer: the application uses Vercel AI Gateway with a conversational model for warm, contextual replies. Deterministic local fallbacks keep the experience useful when the AI service is unavailable.
Safety: medication responses avoid diagnosis, dose changes, or unsafe instructions. Emergency-related messages encourage contacting local emergency services and Esther.
Deployment: hosted on Vercel, with pnpm dependency build permissions configured for reliable production builds. Live demo: https://auracare-omega.vercel.app/

Challenges we ran into

Reliable voice recognition was the hardest part. Browsers behave differently, stop automatically after silence, require permissions, and report interim and final transcripts separately. We kept interim speech visible, preserved final transcripts, restarted sessions when browsers ended them, and added clear microphone-permission feedback.

We also fixed repeated assistant replies, duplicate React keys, and medication-time matching bugs. Getting the tone of safe medical language right took the most iteration: helpful without diagnosing. Finally, we resolved a Vercel deployment failure caused by blocked pnpm build scripts.

Accomplishments that we're proud of

A complete senior-centered experience rather than a generic chatbot
A clear care relationship between Elenor and Esther
Voice interaction with live transcription and automatic sending
Medication reminders with safety-aware guidance
Camera check-in controls and caregiver contact flows
Transparent activity updates that make the system easier to understand
Safety-first behavior without pretending to replace medical professionals
A working, deployed application with reliable production builds

What we learned

Voice UX takes more than turning on a microphone. People need visible transcription, clear listening states, permission guidance, recovery when recognition stops, and a predictable handoff from speaking to responding. For seniors especially, clarity, trust, and caregiver escalation matter more than feature count. Reliable fallbacks are essential because speech and AI services can fail in the real world.

We also learned that a strong senior-focused AI companion should not simply answer questions. It should understand the user's context, communicate calmly, explain what it is doing, provide safe next steps, and involve a trusted caregiver when appropriate.

What's next for AuraCare+

MCP and Alexa+ integration: expose medication, check-in, and caregiver actions through a self-hosted Streamable HTTP MCP server and connect them to Alexa+ tool calling
Proactive ambient care: scheduled check-ins, hydration, meal, and appointment reminders, plus conversation memory with privacy controls
Caregiver tools: secure accounts, multi-senior support, alerts by SMS, email, or push notification, and consent-based wellness dashboards
Safety: fall detection, wearable integrations, location sharing, and stronger emergency workflows
Reach: dedicated speech-to-text, multilingual support, adjustable text size, and improved contrast controls
Personalization: configurable routines, preferred communication style, trusted contacts, and caregiver-approved preferences

AuraCare+ helps seniors feel heard and supported while giving caregivers greater visibility and peace of mind.