
MediVoice is a voice-first medical report companion for people who receive a lab report and do not know what the numbers mean or whom to ask. Many patients wait for a busy son, daughter, or family member to explain their report, leaving them uncertain and anxious. MediVoice lets users upload a medical PDF or report image, then ask questions naturally in English or Hindi. Using AssemblyAI Whisper Streaming for real-time speech-to-text, MediVoice turns spoken questions into report-grounded explanations. It identifies values, units, reference ranges, and printed flags from the uploaded report, then provides a simple, cautious explanation without diagnosing conditions or prescribing treatment. Instead of offering generic advice, MediVoice keeps the conversation connected to the information printed in the user’s own report. Users can ask questions such as “Please explain my report,” “What does this value mean?”, or “Which results are outside the printed range?” The application responds conversationally, helping users understand complex report language without typing long questions or searching through confusing health information online. MediVoice supports English, Hindi, Hinglish, and Hindi transcript normalization, helping users communicate in the language that feels natural to them. When users need next steps, MediVoice can help find nearby doctor or hospital listings and guide them through a simulated appointment-request preview. The experience is designed to be accessible, private, and practical. Uploaded reports are used only for the active conversation, and voice sessions end automatically after a short period. MediVoice is an educational companion, not a replacement for a clinician, diagnosis, prescription, or emergency care.
30 Sep 2026