TrijSign inTrij is an offline-first AI assistant built for community health workers. Capture, listen, and assess — all on the device. No cloud, no compromises, no patient data leaves the phone.
Assessment
Suspected dermatitis
Voice follow-up
"How long has the rash been present?"
Saved offline
3 pending in sync queue
Built for the field
Every piece designed for low-bandwidth, low-light, high-stakes work.
Snap a photo of skin, wounds, or documents. Gemma runs locally on the device — returning condition, urgency, and confidence in seconds.
Iterative follow-ups via on-device speech.
Works on a plane, in a clinic, in a village.
English, French, Swahili, Hindi, Portuguese, Arabic.
Every patient pin lives on the supervisor map.
Live queue, regional analytics, CSV export.
Step away, come back — the voice interview picks up where you left off.
The flow
A workflow tuned for the realities of community health.
Take a photo of the patient's skin, wound, or document with the in-app camera.
Gemma runs locally through WebGPU — no upload, no waiting on connectivity.
Voice follow-up gathers context with function-calling and adapts each question.
Save offline. When you're back online, records sync to your supervisor's dashboard.
Privacy
Images are analyzed locally with Gemma through WebGPU. Only structured, consented assessment records sync to your encrypted backend — and only when you choose.
0 ms
Network latency at inference
6
Supported languages
~2B
On-device model parameters
100%
Of triage runs offline
Trij is free to try. Install it as a PWA and you're ready for the next visit.