GRX Edge is the app clinics install and run on the hardware they already own. It hosts a growing family of diagnostic models — Rad reads chest X-rays today, with more on the way — and a clinician picks the one the case needs. Every read runs on the device.
The app installs freely. Activation requires clinic authentication and approval.
GRX Edge is the app a clinic runs; the models are what actually read a film. Each model reads one kind of image — Rad reads chest X-rays and is first to ship, Mammo is in development for mammography — and a clinician picks the one the case needs, the way you switch models in an app you already use. They share the platform, the on-device footprint, and one engine underneath. Open a model to see what it does.
Flags effusion, consolidation and an enlarged heart on a chest film, region by region — and runs entirely on the device.
The same engine pointed at breast imaging — its own training and viewing objective. What carries over from Rad, and what doesn't, is the work ahead.
One universal model, trained once and shipped everywhere, drifts the moment it meets a population it wasn't shaped for — and quietly loses accuracy exactly where the need is greatest. GRX Edge takes the other path.
A Master Grid holds the known variations of health and pathology for every cell position — and each region has its own. GRX Edge uses GeoLocation to detect where the device is and load the grid built for that population, so a read is measured against the right baseline, not a global average. Move the device and it detects the move and fetches the grid for its new location. Building a grid for a new region takes around 350 films — weeks for a busy centre, not the hundreds of thousands a single global model would demand.
That makes it built for the field. An NGO running mobile screenings, with clinicians moving town to town in a van, gets the local grid at each stop automatically — the tool re-calibrates to each community instead of carrying one model that fits none of them well.
GPS is used strictly for grid selection · accounts are encrypted. Learn more about privacy →A model turns a film into a finding. Getting that model into a clinician's hands, keeping it private, letting them track a patient and pay only for what they use — that's GRX Edge. Here's the shape of it, from getting in to giving back.
Nobody runs GRX Edge without a verified clinic behind them. Inside it, the admin decides who's in — we don't.
One activation binds the app to its device. Films never leave your hardware — only de-identified measurements sync, never the picture.
Pick the reader the case calls for — Rad for a chest film today, others as they ship — from one interface you already understand.
The read runs on the device — no lag on a server, no connection required, low enough power draw to run on a phone.
Link reads to a patient and compare films over time, cell by cell — progression tracking built in, for everyone.
Buy usage credit and spend it on reads, on AIDA, or both — one pool. Contribute verified reads and earn credit back.
Every reader in the family is derived from the same engine, and improved by the same cloud arm. This is the part that makes a lean model possible — and keeps it getting better.
A hybrid transformer fusing self-attention over grid cells with deterministic field physics and rigid attractor mechanics. Its deterministic half is why a model can be inspected rather than probed, trains on hundreds of films instead of hundreds of thousands, and runs on a phone.
Models never learn on the device. AIDA is the cloud arm behind all of them: it ingests verified, de-identified corrections from across the fleet, re-authors each model from that truth, and pushes tiny updates back down. A rare presentation seen once anywhere improves the model everywhere — and no single correction can swing it.
Every read GRX Edge produces is structured data, not just a picture — and GeoRX Lite is the copilot that puts it to work in the clinic. Ask questions of your own patients, follow a case across time, and turn a stack of reads into something you can act on and share.
GeoRX Lite is the light version of GeoRX Research, built right into the app — and it draws on the same in-app credit pool as your reads, so there's nothing separate to buy or manage. The deep, population-scale work lives in GeoRX Research; Lite brings the clinic-sized share of it to where the patient is.
Learn more about GeoRX LiteA small download that runs on the computer you already own. Available from this site only.
GRX Edge is not yet cleared for clinical use. You can download and register now — accounts open when regulatory approval is granted (expected December 2026).