CareBridge.ai — AI Care for loved ones

Inspiration

For millions of immigrant and globally distributed families, caring for aging parents or grandparents from afar is emotionally difficult and logistically inconsistent. You want to check in every day, but time zones, work, classes, and life interruptions make that impossible sometimes. On the other side, parents and grandparents often avoid sharing updates because they do not want to “disturb” their children — and important details get delayed, missed, or forgotten.

We built CareBridge.ai to solve that gap. Not by replacing family, but by supporting family with an empathetic AI layer that keeps care continuous even when communication cannot be.

What it does

CareBridge.ai is an AI-powered family care relay for aging loved ones.

It enables four key workflows:

• AI check-in calls when children cannot call personally
• Smart summaries that capture mood, health, activity, and urgency
• Parent-initiated updates through simple voice or camera interactions
• Urgent escalation to approved family contacts when serious concerns are detected

The AI speaks in a warm, familiar style, supports native-language interaction, and turns conversations into structured updates for family members. Every interaction is triaged into Daily Digest, Care Alert, or Urgent Action, so families know what needs attention now versus later.

In short: CareBridge.ai helps families stay connected, informed, and responsive across distance.

How we built it

We built CareBridge.ai as a full-stack web app using:

• Next.js 16
• React 19
• TypeScript 5
• Tailwind CSS 4

For intelligence and voice, we integrated:

• Google Gemini Live API for real-time multilingual voice conversations
• Gemini 2.0 Flash for summarization, classification, and media analysis
• ElevenLabs Conversational AI for natural voice agent experiences

The product has two primary interfaces:

• A parent-side experience for receiving AI check-in calls and sending updates
• A family-side dashboard for triggering check-ins, reviewing summaries, managing alerts, and configuring care preferences

We used API route handlers to manage call flow, summary generation, alert creation, and settings. For the hackathon demo, we kept persistence lightweight using in-memory and filesystem storage to move fast and focus on the end-to-end experience.

Challenges we ran into

One of the biggest challenges was building something technically impressive without losing emotional trust. This is not just another chatbot — it operates in a deeply personal family context, so tone, clarity, and ethical design mattered a lot.

We also had to solve:

• Urgency detection without overclaiming — identifying emergencies responsibly without making medical diagnoses
• Extreme UX simplicity for older adults — making the parent-side flow intuitive with almost no learning curve
• Real-time orchestration — combining voice AI, multilingual support, summarization, media analysis, and escalation into one seamless product demo
• Balancing empathy and automation — making sure AI feels supportive, not intrusive or replacement-like

Accomplishments that we’re proud of

We are proud that CareBridge.ai feels like a real product with real human value, not just a technical prototype.

Highlights we are especially proud of:

• Built a complete end-to-end demo around a meaningful real-world problem
• Created a dual-sided experience for both families and aging loved ones
• Implemented AI-powered check-ins, smart summaries, multimodal updates, and escalation in one product
• Designed a 3-tier urgency model that makes alerts actionable
• Kept the parent experience minimal, clear, and accessible
• Positioned AI correctly: as a support layer, not a substitute for human care

What we learned

We learned that building AI for caregiving requires more than intelligence — it requires trust, restraint, and empathy.

We also learned:

• The best care tech reduces hesitation and friction rather than demanding behavior change
• Simplicity is critical when designing for older adults
• Multimodal and multilingual AI can be incredibly powerful when grounded in a narrow, meaningful use case
• Escalation systems need thoughtful ethical boundaries, especially in health-adjacent scenarios

Most importantly, we learned that AI can play a valuable role in family care when it strengthens human connection instead of trying to replace it.

What’s next for CareBridge.ai — AI Care for loved ones

We see CareBridge.ai becoming a full remote family care layer.

Next, we want to build:

• Care trend analytics for mood, activity, and adherence over time
• Wearable / IoT integrations for step count, sleep, and heart rate signals
• Shared family circles so siblings can coordinate care together
• Stronger escalation channels including phone calls and richer emergency workflows
• Caregiver handoff notes from local helpers or community caregivers
• Production-grade privacy, auth, and storage for real-world deployment

Our long-term vision is simple:
to help families care across distance with more continuity, less guilt, and faster support when it matters most.

Built With

• elevenlabs
• gemini-2.0-flash
• google-gemini-live-api
• local-filesystem-storage
• lucide-react
• material-symbols
• next.js
• react
• tailwind-css
• typescript