Hydrascan

HydraScan

Inspiration

The project was inspired by a simple gap we kept seeing in the current HydraWave3 workflow: practitioners can capture useful recovery signals, but the path from “I scanned” to “here’s what I should do next” is usually fragmented for the client. We wanted to make that experience feel immediate, supportive, and actionable by combining guided movement capture, recovery scoring, and now curated recovery plans with instructional exercise videos.

What it does

HydraScan is a mobile recovery companion built to help patients turn movement data into clear next steps.

How we built it

We built HydraScan as an iOS app in SwiftUI with a Supabase backend. The app uses QuickPose for guided movement capture and overlays, then structures the scan into a real seven-step onboarding assessment instead of a shallow single-summary blob. We wired the client to real backend services for auth, assessments, recovery intelligence, check-ins, outcomes, and a new recovery-plan system. That recovery-plan subsystem uses patient profile inputs, scan findings, and a safety-reviewed video catalog to surface exercise plans, track completion, and preserve plan history over time. A big part of the work was making the app feel cohesive and premium, so we also redesigned the entire UI to match the Hydrawav3 visual language with a dark-first design system.

Challenges we ran into

One of the biggest challenges was taking something that started as a demo shell and turning it into a real product pipeline without breaking the parts that already worked. QuickPose worked in a verification lab before it worked in the main app flow, so we had to carefully promote that live scanner into the patient capture experience. We also ran into a series of tricky crashes around post-scan results, payload size, and metric labeling that took real debugging on-device to isolate. On the backend side, deployment was not completely straightforward either: we had to repair Supabase migration history drift, deploy new edge functions, and manually apply a large migration safely through the dashboard when the CLI path was blocked.

Accomplishments that we're proud of

What we’re most proud of is that HydraScan is no longer just a proof-of-concept interface. It now behaves like a real end-to-end system: a user can authenticate, complete a real guided scan, persist structured results, see a recovery summary, and move toward a personalized recovery plan built from vetted instructional content.

What we learned

We learned a lot about designing around real-world data completeness, not idealized data. Some scan steps produce rich posture or balance outputs, while others need fallback computation from landmarks when feature-series data is sparse. That forced us to think more carefully about how to represent “complete,” “partial,” and “insufficient signal” instead of pretending every capture is equally clean. We also learned how important it is to separate domains cleanly: recovery-plan adherence should not be mixed with device sessions, and step-level scan data should not be flattened so early that useful clinical detail disappears.

What's next for HydraScan

Our immediate goal is to tighten the loop between the patient's home and the practitioner's clinic. We wanted to implement our patient's daily practices and check-in to the practitioner's side on the actual HydrWav3 dashboard, but we currently don't have access to their source code and backend. We then plan on even recommending where the polarity pads are placed for the practitioner. We have closed the gap for the user, but integrating that with the existing practitioner-side console was our next step so everything is consolidated.