Skip to main content

About the project

Care313: Care that pulls up.

Track: Detroit

Care313 helps Detroit’s mobile health vans bring first-level care to neighborhoods, so emergency rooms stay free for real emergencies. A trainee sits with the patient and Gemma 4 guides the intake. It asks one question at a time in the patient’s language (English, Arabic, Spanish or Bengali) and reads photos of medicine labels, discharge papers and insurance cards. It turns everything into a clear summary. If it spots an emergency sign, it stops and says “Call 911.” A nurse reviews every summary before the doctor sees the patient, and the AI never diagnoses. Coordinators track the vans live on a Detroit map.

Tech stack:

Gemma 4 : the core AI. It runs the multilingual intake, reads photos, checks for emergencies a second time and flags gaps for the reviewer.
Lovable: we built the whole interface with it, including tablet intake, the review console, the doctor view and the live van map.
Open source: the public repo is MIT-licensed, includes reusable intake and review instruction.

Built with

gemmareactopenstreetmaptypescriptvitetailwindcssleaflet.js

AI tools

Coded withLovableClaude CodeClaude Opus 5.5

Prize Categories Entered

  • Deepmind Google

    Best Use of Gemma 4

    Submitted
  • DigitalOcean

    Best Open-Source AI Project

    Submitted
  • LLovable

    Best Lovable Project

    Submitted