Discover how JemmaPass uses a Dual Edge Architecture (Gemma 4 LiteRT-LM & Chrome Built-in AI) to create the world's most resilient clinical data engine — 100% on-device, zero network required.
Forget long forms. Talk to Jemma or point your camera at your medication boxes. JemmaPass uses Google ML Kit Document Scanner (which requires zero camera permissions!) to scan labels and voice intent parsing to understand your symptoms and history instantly.
The data is processed by Gemma 4 running locally via LiteRT. The AI companion maps your inputs to professional clinical standards (SNOMED CT, RxNorm) autonomously, without ever calling a cloud server.
JemmaPass queries its 102 MB Knowledge Base to perform Drug-Drug (DDI) and Drug-Food (DFI) interaction checks. She warns you instantly if a new medication conflicts with your existing profile or allergies.
The Clinical Forge structures your data into an International Patient Summary (IPS). We use advanced clinical pruning to strip heavy metadata, reducing the FHIR bundle size by 70% while keeping it 100% interoperable.
Your profile is ready for the world. Share it via NFC Tap for instant transfer, or display the Triple-Layer QR (Binary, Human, Fallback). It even works from a printed badge or sticker if your phone is dead.
Real-time on-device reasoning demo
.litertlm files hosted directly on jemmapass.com), and an automatic fallback to Chrome Built-in AI (Nano) on the web.