Inspiration
Inspired by the struggles of Thalassemia patients — lack of awareness, late diagnosis, and emergency access to blood donors. Wanted to create a system that detects early health symptoms from voice/text and connects users to help.
What it does
Converts voice to text using speech recognition. Analyzes for health symptoms using NLP. Provides awareness and recommendations on potential health risks. Helps patients find nearby blood donors in emergencies via APIs.
How we built it
Speech-to-text: OpenAI Whisper. NLP: spaCy + custom symptom detection model. Frontend: Electron desktop app + browser extension. Backend: FastAPI + MongoDB. APIs: e-RaktKosh (blood donor info) & hospital databases.
Challenges we ran into
Ensuring privacy with local processing and user consent. Handling false positives in symptom prediction. Simulating APIs where real hospital data wasn't available. Balancing awareness without fear.
Accomplishments that we're proud of
Built a working prototype that listens, analyzes, and connects. Provides both awareness and real-time emergency support. Addresses both preventive care and critical situations.
What we learned
Importance of ethical health tech. Conversations can be a rich source of early health signals. Government/NGO data can amplify real-world impact.
What's next for TalkCare : Voice-Based Health Awareness Recommender
Add multilingual support. Expand to cover more health conditions. Launch as a PWA (mobile/web). Partner with NGOs and hospitals for real deployment.
Log in or sign up for Devpost to join the conversation.