Certify medical AI, from bedside firmware to surgical robots
Engineers see findings at the exact line in VS Code, and a passing model leaves with a signed attestation for auditors.
Medical AI is high risk under the EU AI Act and must also meet device software and risk-management standards such as IEC 62304 and ISO 14971.
Manufacturers need evidence for each software version, produced without exposing patient data or model weights.
Standards in scope
Certification built into the way your teams already work
Findings in the engineer's editor
The CertOne extension for VS Code sends the project for a confidential scan and lists each finding in the Problems panel, at the exact line.
A gate in Azure ML
Training pipelines call CertOne before packaging, and only a passing model version moves on.
A signed attestation for auditors
A passing verdict yields a signed compliance attestation, bound to the exact snapshot and revoked automatically if a covered component changes.
The pipeline gate certifies each version.
At the exact line, in the editor.
A signed PDF for auditors.
Recorded on the live CertOne platform.
Embedded firmware, a pure-C int8 1-D CNN for an Arm Cortex-A72 patient monitor.
A tumour-tile classifier (SqueezeNet, ONNX) for digital pathology.
Vessel-proximity warning for a surgical robot (LR-ASPP segmentation, ONNX).
See CertOne on your industry's systems
Register to watch the full walkthroughs, recorded on the live platform.