AuraHealth

AuraHealth was inspired by the "care gap" in healthcare, where patients often feel lost during post-consultation recovery while doctors are too overwhelmed to proactively check in on every individual. Our platform bridges this gap by automating follow-ups using a voice-first AI agent that calls patients, triages their symptoms via Gemini 2.5 Flash, and escalates red flags to a doctor dashboard in real-time. We built the system using a FastAPI backend integrated with Twilio for telephony, Google Cloud for STT/TTS, and Pinecone for RAG-powered clinical context. One of our biggest challenges was optimizing the audio pipeline to eliminate latency, ensuring the AI maintains a natural, empathetic, and responsive conversational flow. We are incredibly proud of achieving a seamless voice interface that successfully translates complex medical data from PDFs into personalized, patient-friendly check-ins. Through this process, we learned the importance of "Small Model, Big Reasoning" for real-time clinical applications and how to manage high-concurrency audio streams effectively. Moving forward, we plan to expand AuraHealth with multilingual support, emotion detection, and direct integrations with major EHR systems like Epic and Cerner.