PatientVoice

PatientVoice helps adults communicate with healthcare staff when speaking or composing a message is difficult. A patient can select a phrase, type, or record a short message. Meta Muse Voice Transcribe turns recorded speech into an editable draft. The interface keeps the interaction simple: Record, Rewrite, Send. Optional Meta Muse Spark rewriting temporarily replaces the text box with a suggestion. Accept returns the revised wording to the editor; Reject restores the original. Only Send delivers the message. Staff receive the patient’s original input alongside accepted wording and can explicitly acknowledge receipt. The patient sees that acknowledgment, closing the loop rather than leaving them unsure whether their message was received. The prototype includes separate patient and staff views, original-text preservation, protection against stale suggestions and conflicting browser tabs, duplicate-safe retries, and direct messaging when AI is unavailable. PatientVoice is communication assistance—not a system for diagnosing, interpreting symptoms, recommending treatment, or replacing emergency systems or qualified interpreters. Demonstrations use synthetic data; the product is not clinically validated. PatientVoice is the focused, accessibility-oriented prototype developed from our original ClinicAI application idea. All demonstrated product functionality was built during the hackathon, with implementation history and development provenance documented in the public repository.