Rehably

Inspiration

We wanted older adults and people managing chronic conditions to take responsibility and acountabilty over their physical therapy at home. Clinical research shows that adherence drops when patients lack real-time feedback, so we set out to build a coach with measurable progress. The vision: make high quality rehab support as accessible as opening a browser.

What it does

Rehably provides a web-based rehabilitation companion that:

Guides users through condition-specific routines such as sit-to-stand and arthritis-friendly exercises.
Tracks repetitions, sets, range-of-motion angles, and time-under-tension with real-time feedback.
Generates session summaries and printable reports so patients can share progress with clinicians.
Offers adaptive tips and encouragement to keep users motivated between in-person visits.

How we built it

Front end: HTML, CSS, and vanilla JavaScript for lightweight, offline-friendly pages across index.html, conditions.html, and condition-specific exercise flows.
Motion intelligence: WebGL and wasm-powered pose estimation to measure joint angles and count clean reps directly in the browser, minimizing latency and preserving privacy.
Data handling: Client-side storage for quick session caching and structured exports for clinician review.
Design system: Reusable components (cards, accordions, CTA buttons, timers) styled consistently via styles.css to keep the experience accessible and responsive.

Challenges we ran into

Calibrating pose detection to handle diverse camera angles and lighting without frustrating false negatives.
Creating UX copy that is supportive but clear for non-technical users(seniors).

Accomplishments that we're proud of

A fully client-side experience that protects privacy while delivering measurable feedback for clients as well as doctors/care givers.
Report generation that clinicians can quickly scan, reducing friction during follow-up visits.
An accessible interface with large controls, contrast-aware palettes, and concise instructions tailored to seniors.

What we learned

Small UX touches (progress rings, gentle toasts, concise “why it matters” blurbs) dramatically improve adherence.
Running pose estimation in-browser is viable with careful model selection and throttled rendering.
Clear error states and calibration tips reduce drop-off during onboarding.