What Is Aidoc? The aiOS Platform, CARE Model and FDA Clearances Explained

The short answer: Aidoc is a US-focused clinical AI company whose software reads CT and X-ray studies as they arrive, flags suspected time-critical findings and pushes a notification to the right team so those cases move up the queue. Its products run on a platform the company calls aiOS, powered by a foundation model it calls CARE, short for Clinical AI Reasoning Engine. In January 2026 the FDA cleared an Aidoc solution covering 14 body CT indications in a single workflow. It is a triage and notification product: it does not write your reports, and the radiologist reviews and signs every study.

Updated August 2026. Clearance lists in this market change quickly, so confirm the current indication set with the vendor before planning around it.

Aidoc comes up in almost every radiology AI evaluation in the US, often as the first name on the list, and it is frequently misunderstood in both directions. Some groups assume it is a general-purpose radiology AI that reads everything. Others assume it is a single stroke tool. Neither is right. This guide explains what the product actually is, what the January 2026 clearance changed, and where it stops.

What is Aidoc?

Aidoc is a clinical AI vendor, not a PACS, a viewer or a reporting platform. Its software sits alongside the systems you already run: studies flow to it from your archive, its models analyze them, and when something suspicious turns up it raises a flag and notifies whoever needs to act. The radiologist still reads the study and still writes the report. What changes is the order cases are read in, and whether a quietly urgent study sits in the queue for two hours before anyone opens it.

The company's framing is that it operates at the health-system level rather than the department level. That is a real distinction when you are buying. Aidoc is most often sold to a system standardizing AI across many sites and many pathologies, rather than to a single practice buying one tool for one problem.

What does Aidoc do?

Three things, in sequence. It analyzes studies automatically as they arrive, without a radiologist asking it to. It flags suspected findings on the pathologies it is cleared for. And it notifies, which is the part that changes behavior: the alert reaches the people who can act, so a suspected finding on a study nobody has opened yet still starts a clock.

That notification step is why triage vendors describe their value in minutes rather than in detections. If a case would have been read at the same time anyway, an accurate flag has not changed the outcome. The value appears when a study that would have waited gets read early.

Is Aidoc FDA approved?

Cleared, not approved, and the distinction is worth getting right because it changes what the claim means. Radiology AI of this kind goes through FDA clearance pathways such as 510(k) or De Novo, which establish that a device is substantially equivalent to a predicate or that its risk is adequately controlled. "FDA approved" describes a different and more demanding pathway used for higher-risk devices. Aidoc holds clearances, and a vendor telling you it is "FDA approved" is a signal to read the actual clearance letter.

The clearance that matters most is recent. On January 26, 2026 the FDA cleared what Aidoc describes as the first comprehensive triage solution built on a healthcare foundation model, bringing 11 newly cleared indications together with 3 previously cleared ones into a single body CT workflow, 14 in total. Named indications include spleen, liver and kidney injury, pelvic fracture and intestinal ischemia. In the pivotal study submitted to the FDA, the company reports mean sensitivity of 97% and mean specificity of 98%, reaching 98.5% and 99.7% at the top of the range, and describes roughly an order-of-magnitude reduction in false alerts compared with single-condition tools. Those are Aidoc's own published figures for Aidoc's own product, and they are worth asking a vendor to reproduce against your case mix rather than accepting from a slide.

What is aiOS?

aiOS is the layer Aidoc runs its clinical solutions on. Rather than integrating each model separately into your PACS, RIS and messaging, a system integrates the platform once and then turns individual cleared solutions on within it. Aidoc states that more than 100 million patient cases have been analyzed through it.

The buying logic follows from that. If you expect to run one AI tool for one pathway, a platform layer is overhead you do not need. If you expect to run six over five years, the integration and governance savings are the whole argument, and it is the argument Aidoc leads with.

What is CARE?

CARE, which Aidoc expands as Clinical AI Reasoning Engine, is the company's own foundation model, and it is what the January 2026 clearance was built on. The practical difference from the older approach is architectural. Traditionally each indication meant a separate narrow model, separately trained, separately cleared and separately generating its own false alerts. A foundation model trained across imaging is meant to support many indications at once, which is how 11 indications arrived in a single clearance instead of eleven.

The company has said it expects the CARE roadmap to extend across all CT and X-ray workflows over roughly the following 18 months. Treat a roadmap as a roadmap: useful for understanding direction, not something to write into a contract.

Does Aidoc write the radiology report?

Not today. Aidoc is a triage and notification product, so the report is still written wherever you write reports now, whether that is PowerScribe, DeepHealth Reporting Pro or something else. In June 2026 the FDA granted Aidoc a Breakthrough Device Designation for First Read, which analyzes chest radiographs and generates preliminary report text. A Breakthrough designation is a program that expedites review; it is not a clearance and it is not a shipping product, so a group planning around report drafting today should not count on it.

How much does Aidoc cost?

Aidoc does not publish list pricing, and nobody serious in this market does. Pricing is quoted per organization and normally shaped by which cleared solutions you enable, study volume, number of sites and the integration work involved. Any single figure circulating online is somebody else's negotiated deal.

The number worth pinning down is not the license, it is the fully loaded first-year total plus what the subscription does at renewal three years out. Interface work with your PACS and RIS, validation against your own case mix, and the security review each new vendor triggers are all real costs that rarely appear on the first quote. Systems that end up running several AI subscriptions alongside their other software eventually want a single view of what every one of those recurring subscriptions is actually costing across the year, because per-vendor renewal dates have a way of drifting apart until nobody owns the total.

What is the difference between Aidoc and Viz.ai?

Breadth versus coordination. Aidoc covers more conditions through one platform, optimizing how much of the acute reading list is watched. Viz.ai is built around the stroke pathway and alerts the specialist team directly on mobile, optimizing how fast the right person starts moving. They overlap on stroke and diverge almost everywhere else, and RapidAI is a third answer again, going deeper into stroke imaging itself.

Because that three-way choice is the one most US groups actually face, we built a neutral side-by-side table of all three, covering FDA clearances, non-contrast CT, Medicare reimbursement and where each one stops: Aidoc vs Viz.ai vs RapidAI. One detail from it is worth repeating here, because it surprises people: Viz LVO holds the first New Technology Add-on Payment CMS ever granted to AI software, and neither Aidoc nor RapidAI has a published equivalent.

Where Aidoc stops

Every triage platform draws a boundary around the urgent slice of the worklist, and Aidoc draws a wider one than most. The boundary is still there. Routine studies, the bulk of any reading list, are not what a triage clearance covers, and the report gets written the same way it always did. Groups usually discover this about six months in, when triage is working well and the shift still ends at the same time.

That is the gap Radiological.ai is built for. It flags suspected findings for your review, prioritizes the worklist so urgent studies surface first, and drafts the structured report into your template before you sit down, across X-ray, CT and MRI in one pane, for routine studies as well as urgent ones. It is decision support rather than a diagnosis, and the radiologist reviews, edits and signs every study. If you are evaluating Aidoc specifically, our Aidoc alternative comparison puts the two side by side, the radiology worklist prioritization page explains the queue mechanics underneath both, and the radiology AI alternatives hub groups every vendor comparison by what you would actually be replacing. The illustrative sample above shows how a drafted report reads.