The compliance layer for clinical AI.
Hospitals are deploying AI faster than they can govern it. Orinyx is the layer that checks every recommendation against the sources that actually matter — FDA drug labeling, clinical pharmacology guidelines, and federal medication safety guidance — and writes an audit log your boards, regulators, and lawyers can read.
How verification happens
Sit in the path
Orinyx installs as one layer between the AI tool and the EHR. No clinician workflow change. No new screen to learn.
Check the sources, not the model
Every recommendation is checked against FDA drug labeling, clinical pharmacology, and federal medication safety guidance. We don't trust the model — we verify the claim.
Flag, attach source, log
Unsupported recommendations are flagged Minor, Moderate, or Major. Supported ones pass through with the source attached. Every output — flagged or not — is logged.
Flag severity
Three levels. Same audit trail.
Documentation gap
AI suggested "ibuprofen 400mg PRN for headache" without noting the patient's recent GI bleed history in the chart. Drug is appropriate; documentation needs a note.
Dose outside label
AI recommended sertraline 150mg/day for an elderly patient. FDA label suggests starting at 25–50mg in geriatric patients. Recommendation flagged for review before signing.
Contraindicated combination
AI recommended apixaban + clarithromycin without dose reduction. FDA label requires 50% dose reduction for this combination due to CYP3A4 interaction. Blocked from signing.
The audit log
The same log answers three different rooms.
Quarterly proof that AI deployed across the system is monitored, flagged, and bounded — with rates and trends per service line.
Every recommendation, every check, every source citation. Reviewable on demand for ONC, OCR, and state-level audits.
When a case lands, the log answers: what did the AI say, what did Orinyx flag, what did the clinician see, what did they sign.
Join the platform waitlist.
We're working with a small number of US hospitals during beta. Tell us where you are and what AI you're using.