Inspiration

The power goes out. The patients don’t disappear.

Our demo begins with a truck hitting a power pole, leaving Dr. Helen Miller’s cloud-dependent practice at a standstill. The information needed for a consultation is suddenly out of reach but the patient is still sitting across from the doctor.

Paper can keep a visit moving, but it cannot bring back patient history and it leaves more work when systems recover.

We asked: what if the phone in the doctor’s pocket could keep the visit going?

What it does

SteadyCare connects three Apple devices into one clinic workflow:

  • iPad: Patients speak or type their check-in answers, then review and confirm.
  • Mac mini: The clinic hub holds records and supports consultation transcription and draft notes.
  • iPhone: GPs review patient context, ask questions with cited answers and approve notes.

Online, users can toggle between on-device intelligence and external models, including scx.ai for transcription. Offline, SteadyCare uses only on-device intelligence.

Before an outage, the iPhone saves a day pack of today’s queue, patient briefs and record excerpts. If the Mac becomes unreachable, that context remains available. Compatible iPhones can answer questions from saved excerpts and draft notes using Apple Intelligence. Approved work saves back when the connection returns.

How we built it

We built SteadyCare with Swift and SwiftUI, combining Apple’s Foundation Models, Speech and Translation with scx.ai for optional online transcription.

To fit the on-device model’s limited context window, we retrieve relevant excerpts and summarize in stages. Answers link to sources and dates, and doctors review and approve generated notes.

Challenges and lessons

Our biggest challenge was making offline continuity dependable. Day packs must be saved beforehand, and new check-ins still require the Mac.

We learned that resilience starts before a connection fails and that useful clinical AI needs checkable sources, clear limits and human judgment.

What’s next

SteadyCare is our MVP. We plan to add ElevenLabs to the online version only, giving patients a natural voice conversation at the start of check-in while retaining on-device speech for offline use.

We’ll also integrate directly with practice record systems, test the workflow with clinicians and refine the experience based on their feedback.

SteadyCare: care that keeps going, even when the clinic goes offline.

Built With

  • apple-intelligence
  • core-ml
  • coreai
  • foundation-models
  • healthkit
  • ios
  • ipados
  • macos
  • natural-language
  • scx.ai
  • speech
  • swift
  • swiftui
  • translation
  • vision
  • xcode
Share this project:

Updates

Submission history