Established MMXXIV

Medical imaging, made differentiable.

We build embedded AI segmentation SDKs that medical-imaging OEMs license and ship inside their own viewers and planning suites — production C++ that runs in-process on the workstation GPU, with no cloud, no service to operate, and a regulatory dossier that plugs into your file.

Mathematical Precision • Engineering Rigor
  • For medical-imaging OEMs

    Embed segmentation. Skip years of build.

    License a binary C++ SDK that runs in-process in your viewer or planning suite — a library in your binary, not a service to run. We deliver the segmentation engine and the component-level regulatory dossier; you carry the product and the clinical claim.

    See the product
  • For investors

    A specialist company with a beachhead and a horizon.

    A specialist company with a beachhead in CBCT image segmentation — dental today, more CBCT anatomies next — and a long-running focus on the foundations of medical-imaging segmentation that predates the current deep-learning wave. European AI company headquartered in Italy, senior engineering team, B2B2C SDK economics — building the kind of imaging-AI company that scales by partnership, not by patient throughput.

    Read about the company
  • For clinical partners and KOLs

    Imaging engineers who take provenance seriously.

    Validation studies, dataset licensing, and clinical-advisor relationships, on retrospective anonymised data, on terms that respect the clinic's role and the patient's privacy.

    Start a conversation
Today's product

More than 70 classes of dental CBCT segmentation, embedded.

Today the product is CBCT image segmentation, beginning with more than 70 classes of dental cone-beam CT. As the catalogue grows, the same SDK foundation extends to adjacent CBCT anatomies, while our work stays on the foundations of medical-imaging segmentation. Segmentation only; clinical interpretation is the OEM's and the clinician's responsibility.

  • Runs in-process in your binary — no container, no local service, no cloud round-trip. A full clinical CBCT in less than a minute of inference on a single, modest workstation GPU (RTX A4000) — no high-end card, no separate server or GPU cluster, and faster still on newer hardware.

  • Production C++ binary SDK for Windows and Linux.

  • Fully encapsulated public API. No third-party headers in your build.

  • Regulatory dossier package (SOUP, model card, validation summary, IFU template) that plugs into your CE MDR or FDA 510(k) technical file.

  • Model IP protected: encrypted, licence-bound model artifacts, and a security posture assessed against IEC 62443, ISO/IEC 27001, and the EU Cyber Resilience Act.

  • Reproducible by policy: every shipped model is uniquely versioned and traceable to a sealed, quality-gated training run.

Built for OEM economics, not for research demos.

  • From first principles.

    PhD-level applied mathematics on the inside, a clean binary on the outside. We work from the mathematics of medical volumes — geometry, numerical analysis, optimisation theory — up to the C++ that ships in the customer's installer. We do not glue libraries together and hope.

  • Regulatory leverage.

    Our component-level dossier package drops into the OEM's CE MDR or FDA 510(k) file. We deliver the software-component documentation; the OEM carries the clinical claim. We are a component vendor, not a finished-device manufacturer — and we engineer accordingly.

  • No surprises in the customer's binary.

    Embedded utility dependencies, encapsulated public API, signed Windows installers, reproducible Linux builds. The only thing the customer sees is a stable C++ contract and a sealed model artifact.

Beyond the SDK

Partnerships for complex AI/ML problems and critical system software.

For a small number of organisations that need the same depth of engineering that built the SDK — applied to a problem of their own — we open a focused partnership track. Senior team only, narrow scope, deliverables we ship rather than slides we present.

Learn more

Build medical imaging hardware or planning software?

If you are tired of evaluating research notebooks dressed up as products, we should talk.