HackStack

Project Info

Winner

Claude: Best Use of Claude

Elda AI

Devpost

This project did not submit a demo video on Devpost.

Inspiration

When my grandmother was diagnosed with dementia, we watched her memory fade and her independence slowly disappear. Simple routines like taking medicine or remembering meals became daily struggles. That experience made us realize how deeply technology fails the people who need it most. EldaAI was born from the desire to build an AI that could understand, care, and detect before things get worse.

What it does

EldaAI is an AI-powered eldercare platform that transforms caregiving from reminders to real-time intelligence. It listens, learns, and detects using Claude to understand speech and emotion, Letta to learn daily behavior patterns, and Chroma to detect changes over time. EldaAI automatically generates daily and monthly health summaries for doctors and caregivers, highlighting medication compliance, mood shifts, and potential cognitive decline.

How we built it

Backend: FastAPI with PostgreSQL, JWT authentication, and APScheduler for smart reminders. AI Layer: Claude for intent and sentiment analysis, Letta for behavioral memory, Chroma for semantic pattern detection. Mobile App: Built with React Native + Expo for voice-first interaction. Caregiver Dashboard: Built with Next.js + Tailwind for real-time monitoring, insights, and alerts. Communication: Twilio for voice calls, Firebase for push notifications. Deployment: Railway (backend), Vercel (dashboard).

Challenges we ran into

Making voice-based AI interactions natural for elderly users with limited tech familiarity. Training Claude to detect subtle emotional cues and health-related intent. Integrating multiple AI services (Claude, Letta, Chroma) seamlessly while staying within API limits. Designing a UI that feels simple enough for 80-year-olds but powerful enough for doctors.

Accomplishments we're proud of

Built a fully functional prototype Mobile App for Elder People and Care Giver web app in 2 days that integrates Claude, Letta, and Chroma. Successfully generated real-time AI health summaries from daily conversations. Designed a caregiver dashboard that visualizes daily mood and medication adherence. Created a system that bridges empathy, AI, and healthcare — something our team truly believes in. Building human-centric AI means designing for emotion, not just accuracy. Simplicity in UX is more powerful than complexity in features — especially for elders. Prompt engineering and data context are key to getting reliable results from LLMs like Claude. Collaboration across AI, design, and healthcare perspectives can create deeply impactful solutions.

What's next

for EldaAI Integrate with wearable health data (heart rate, sleep, activity) for early disease detection. Enable multi-language and cultural personalization for global accessibility. Expand caregiver tools to include doctor collaboration and long-term pattern analytics. Long-term vision: Let people “train” their AI throughout their lives so when they age, their AI already knows them deeply, acting as a personal caregiver and memory companion.

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • AnthropicIn code
  • CIn code
  • CSSIn code
  • FastAPIIn code
  • HTMLIn code
  • JavaIn code
  • JavaScriptIn code
  • Next.jsIn code
  • PythonIn code
  • ReactIn code
  • Tailwind CSSIn code
  • TypeScriptIn code

12 of 12 appear in the indexed code.

AI coding agents

  • Claude CodeConfig · Commits

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

1.9 MB

Source files

271

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars