Product
Built for the nurse, the family and the home.
A workspace where agents prepare the work and clinicians decide, an app that keeps families close to the care, and a vision model that notices when a patient falls.
Clinician workspaceLive in part
Agents prepare the work. Clinicians decide.
Overnight, agents draft visit notes, check assessments and book follow-ups. In the morning the clinician reviews what they prepared and signs it, and through the day the dispatch agent keeps her route current.
Synthetic patients throughout. Notes, signing and the copilot run in today's product; the agent queue and dispatch are being built.
Family appBuilding
Family can see the care as it happens
Patients choose who can follow their care and what each person sees. Family get a plain summary when the nurse leaves, can message her directly, and hear first if something goes wrong.
The family app is in development. Eleanor Brooks and her family are synthetic.
Fall detectionTraining
A vision model that notices a fall
We are fine-tuning an open-source vision-language model on falls at home. It follows a person's pose frame by frame, confirms a fall when they stay down, and starts the response: the nurse on call, the family and the schedule.
Real footage: a staged fall from the GMDCSA-24 research dataset (Alam et al., 2024, MIT License). The keypoints, box and metrics come from MediaPipe Pose run on that footage. The scene descriptions stand in for our fine-tuned model, which is still training, and the response is illustrative.
The next step
Three technologies.
One connected care experience.
One suite of software that bridges home health and nursing home care, expanding what's possible for patients cared for at home.