Immersia

Inspiration

Dementia does more than degrade memory; it disrupts a patient's spatial and temporal mapping, leading to immense anxiety, isolation, and a loss of identity. While clinical Reminiscence Therapy (using past experiences to evoke joy and comfort) is highly effective at reducing these symptoms, it places a heavy, constant burden on caregivers. We wanted to build a "cognitive prosthetic"—a digital tool that delays the inevitable progression of the condition by actively anchoring patients to their reality, bringing vital comfort to both patients and their families.

What it does

Immersia is a Digital Therapeutic (DTx) platform built to deliver immediate, personalized reminiscence therapy. Designed with a zero-learning-curve interface for elderly users, it provides caregivers with a dashboard to launch two distinct, high-impact therapy modes:

Slideshow Therapy: A soothing visual module that cycles through happy family photographs accompanied by the patient's favorite music, designed to quickly reduce anxiety.

Voice Memory Video: A deeply immersive nostalgia feature that plays generated spatial memory videos, narrated by hyper-realistic, AI-cloned voices of the patient's loved ones (like a spouse or child).

How we built it

We built Immersia as a Single Page Application (SPA) utilizing a modern, highly accessible tech stack:

Frontend: React and Vite, utilizing React State to seamlessly route between the caregiver dashboard and the therapy modules without confusing page reloads.

UI/UX: Tailwind CSS and Lucide-React. We engineered a strict clinical design system (Slate Off-White, Deep Navy, and Warm Coral) featuring massive hit areas, soft drop shadows, and high contrast specifically tailored for cognitive decline accessibility.

Voice AI Engine: We integrated the ElevenLabs Voice Cloning API to generate the hyper-realistic, asynchronous audio anchors of family members.

AI Architecture: We used Gemini as our anti-gravity harness to architect the frontend structure, enforce accessibility constraints, and generate the UI components.

Challenges we ran into

Our biggest challenge was the pivot in UX. Initially, we wanted to build a complex 360-degree navigable spatial environment. However, we quickly realized that complex navigation is the enemy of dementia care. We had to ruthlessly strip away our own complex features to create a flat, zero-learning-curve interface. Furthermore, syncing the emotional weight of the ElevenLabs cloned audio with the HTML5 visual media required careful state management in React to ensure the therapy didn't feel robotic or disjointed.

Accomplishments that we're proud of

We are incredibly proud of building a deeply empathetic, non-pharmacological healthcare tool in just 3 hours. We successfully integrated cutting-edge AI (ElevenLabs) not just as a gimmick, but as a core medical feature that turns standard video playback into a deeply personal, emotional anchor. We are also proud of our strict adherence to accessibility guidelines, resulting in an interface that a patient can actually use without frustration.

What we learned

We learned that in Digital Therapeutics, UX is the medicine. A confusing interface can trigger the exact anxiety we are trying to cure. We also learned the profound psychological power of audio—visuals help set the scene, but hearing the exact voice of a loved one is what truly unlocks deep-seated memories and emotional comfort.

What's next for Immersia

Dynamic Generation: Integrating real-time image-to-video AI so caregivers can dynamically generate memory scenes from single old photographs.

Emotional Tuning: Utilizing advanced ElevenLabs parameters to dynamically adjust the tone of the cloned voice (e.g., more soothing when the patient's heart rate or anxiety is high).

Wearable Integration: Connecting Immersia to smartwatches to automatically trigger the 'Slideshow Therapy' and favorite music when a spike in the patient's heart rate indicates a panic or wandering episode.