CareHandoff

Inspiration

Nonprofit organizations face high volunteer and staff turnover with limited resources for structured handoffs. When people leave, critical knowledge leaves with them — often undocumented.

CareHandoff turns voice into structured records. Speak for 30 seconds, and the system captures and organizes everything for the next person. No forms, no training.

Built for animal shelters where the problem is most urgent, but designed to scale across any nonprofit where people rotate and knowledge gets lost.

What it does

CareHandoff is a seamless, voice-first dashboard. Volunteers simply speak their updates after checking on an animal. The app automatically transcribes the audio, extracts key alerts, and generates a structured care log.

Highly Accessible & Cost-Effective: Volunteers can speak in their native languages, and the app unifies the output in English so the entire team shares a readable, common record. Thanks to our optimized AI pipeline, processing a complete check-in costs less than 1 cent.

How we built it

🎙️ Voice-to-Insights: Multilingual voice input (Korean, Mandarin, Punjabi, etc.) with automatic English transcription for team-wide accessibility.

🧠 AI Task Extraction: Automatically pulls active alerts and pending tasks from the spoken update.

🔍 Interactive Transcript: Hover over any alert or task to instantly highlight the source phrase in the transcript.

🐾 Shelter Dashboard: A clean, centralized view to browse all animals, care histories, and to-do lists at a glance.

🛠 Tech Stack Frontend: Next.js, TypeScript, Tailwind CSS, shadcn/ui Backend: FastAPI (Python) AI: OpenAI Whisper (Speech-to-Text) + GPT-4o-mini (Data Extraction)

Challenges we ran into

One of the biggest challenges was deciding on the idea in the first place. Since the nonprofit case involved so many different problems, it was difficult to determine which issue we should focus on and at what scale we could realistically solve it. It was also challenging to bring together all of our team members’ ideas into one clear direction. In the end, we decided to keep the core concept of voice activation and think carefully about where it could create the most value. Through team discussions and mentor feedback, we narrowed the idea down to a handoff-focused solution.

Accomplishments that we're proud of

We are proud that we turned a broad, complex problem into a more specific solution with real impact. Every team member contributed well in their own role, and we worked in a very synchronized way. We are also proud that the voice activation worked accurately, including multilingual recognition, and that the AI summarization and other core features functioned successfully.

What we learned

We learned a lot about how to implement and connect AI features in a working product. In particular, we gained experience integrating the frontend and backend smoothly and making sure the whole system worked end to end. We also learned how to use the OpenAI API in a practical project setting.

What's next for CareHandoff

Our next steps are to support multilingual output, add a feature that lets users see summaries at a glance from the main page, and optimize the product so it can be easily adapted for other nonprofit organizations as well. We also want to think more carefully about alternative input methods for situations where speaking out loud is difficult, such as quiet environments or contexts where privacy is important.