Software that learns each clinician's own words for how they like to read a scan — sharper, brighter, show the bone — and adapts the display to that person over time.
One brain · any modality · learns each reader
— and adapts the display to that person over time
Dictation is standard. Voice-driven image navigation is arriving, and there's now a 2025 proof-of-concept for language-model voice control inside the MRI scanner room. Real progress — but it's the same display, behaving the same way, for every reader who uses it.
Hanging protocols and window presets arrange and contrast the image by static, pre-set rules an admin configured. The settings mean the same thing for everyone. Your "sharper" and my "sharper" get the exact same window. The tool never gets to know you.
The same word means something different to every reader. Our software learns what you mean by "sharper," remembers it, and tunes the display to you. The more you read, the more it becomes yours. That is the missing piece.
Radiology has long accepted that how a study is laid out and windowed is personal. Hanging protocols, in the words of the imaging-informatics field, vary by "modality, body part, department, personal preference, and even training" — one radiologist wants the chest view on the left, another on the right. But today that preference is set up by hand, as static rules.
The first, demonstrable product: a voice-driven CT/MRI viewer that windows and switches presets from plain language and learns each reader. It runs today on real, open imaging datasets — the clearest, simplest place to prove the concept in medicine.
The same software learns a clinician's preference for almost any imaging or image-guided system they read or operate. The CT/MRI viewer is where we start; this is the wider imaging world it fits.
This personalizes how an image is displayed and learns a reader's preferences. It does not detect, measure, diagnose, or interpret findings, and its output is never represented as diagnostic. Every change is written to an audit log. It is built research-first, with the cleanest regulatory posture in mind.
Imaging volumes keep rising, AI is pouring into the reading room, and the fastest growth is in the imaging software and AI layer — exactly where our personalization lives.
Sources: Grand View Research (AI in medical imaging); Persistence Market Research (imaging informatics); Coherent Market Insights (PACS & RIS).
Imaging vendors and health systems pay recurring for software per reading seat. Recurring costs require recurring revenue, so it's built as a per-seat license folded into the product.
| Scenario | Reading seats | Our revenue / year |
|---|---|---|
| Early | 2,000 | $4.8M |
| Growing | 10,000 | $24M |
| Strong | 30,000 | $72M |
| At scale | 60,000 | $144M |
Illustrative at $200 per reading seat per month — final pricing is set with each partner. Reading seats span every modality and geography; even 60,000 is a fraction of the global reading workforce, and every seat pays each month it runs.
A provisional patent application has been filed on the core method, with broad cross-domain claims in progress. The method is held as a trade secret and is not disclosed here. Full details are shared only under a signed NDA.
For the first time, the image can learn each clinician's own language for how they read — and become truly theirs, across every modality. It works today, on real scans. We're building the company that brings it to the imaging world.
Confidential · Research prototype, not a diagnostic device · Do not share without a signed NDA