One AI-native platform across endoscopy, pathology and radiology.

Some departments want everything to stay inside their own walls; others want the reach of the cloud for review, teaching or a second site. CloudKites does not pick the winner. A pod can run fully on-prem, keeping the archive and the information system inside the facility. Or it can run hybrid — an on-prem pod paired with a cloud pod — where the two stay in sync asynchronously, so the edge keeps working through interruptions and the studies become reachable wherever they are needed. The arrangement is yours, and it can change as you do.

The same flexibility runs the other way too — in scale. The architecture flexes from a single small clinic to a medium-to-large hospital without changing how it behaves. A small anatomical-pathology lab can run PathoPod as its LIS; a busy radiology department can run TomoPod as a full AI-native RIS and PACS; an endoscopy unit can run EndoPod as its EIS. You add capacity and modalities as you grow, and the platform you knew on day one is the platform you keep.

It would have been easier to build three separate products, each with its own habits, its own quirks, its own definition of what counts as safe. Plenty of software grows that way, one team at a time, until the seams show. CloudKites chose the harder path on purpose: one AI-native foundation, expressed as an edge tool and a pod for each modality. The endoscopist, the pathologist and the radiologist see tools shaped to their craft — EndoEdge and EndoPod, PathoEdge and PathoPod, TomoEdge and TomoPod — but they are standing on the same ground.

The payoff is that the model does not fragment. The edge assists in real time and captures; the pod is the information system — EIS, LIS or RIS — encapsulating studies as DICOM and aligning to HL7 with a HIPAA-aware posture. Deployment is on-prem or hybrid by choice. AI plugs in gradually and is customisable per customer. And the clinician stays in the loop, reviewing and signing the semi-automated report. When a department adds a second or third modality, it is not learning a second or third way of working. It already knows this one.

There is a quieter benefit too. A shared foundation means a model that helps in one modality can be extended to the next, and a standard met in one place is a standard met everywhere. The platform gets better as a whole rather than in disconnected pieces — and the clinician, who never asked to think about any of this, simply finds that the edge tool assists and the pod behaves the same way wherever they meet it.