NeroLens is a supportive app for people with Parkinsons that uses Google Gemini 3 as its core intelligence. On the home screen a Gemini-powered chatbot answers questions, collects symptoms, and guides users to the specific tests. The smile analysis feature lets users upload or capture a selfie; Gemini analyzes facial landmarks and expression dynamics to produce an expressivity score, a confidence level, and a short plain-language explanation of what changed from prior tests. The MRI review feature accepts uploaded brain scans, where Gemini helps identify and highlight visual patterns, provides annotated overlays for key regions, and offers a stage suggestion together with notes that a clinician could review. The handwriting module provides a canvas for drawing a spiral; Gemini extracts tremor amplitude, smoothness, and pen-pressure proxies, quantifies micrographia-like changes, and returns both a numeric tremor index and an annotated spiral visualization. The exercise module asks five targeted questions, then uses Gemini to synthesize the answers, estimate a likely stage, and recommend a tailored, safety-minded exercise routine with step-by-step instructions and a short example clip. All analyses include brief explainable reasoning rather than opaque labels, and are framed as supportive guidance, not medical diagnoses. For reliability and testing, developers can swap the Gemini API key in settings to change endpoints or use fallback keys, while the app sends secure requests to a backend that orchestrates Gemini calls and stores encrypted data. Full documentation, an architectural diagram, and links to the demo and code are included in the project repo.
Log in or sign up for Devpost to join the conversation.