Registration at the front desk, scans pushed straight from the machine, an AI second reader beside the radiologist, and the report signed on the platform — with a link of the study you can send to the referring doctor and the patient.
Every role in the centre works on the same record — the booking made at the desk is the study the machine scans, the case the radiologist reads, and the report the referrer receives.
Name, procedure, referring doctor, slot — the booking appears on every worklist and is published as a DICOM Modality Worklist, so it's on the scanner before the patient is.
The technologist selects the patient on the console itself. Name, ID and accession come from RadX — nothing is typed at the machine, nothing can mismatch.
The study is pushed over the DICOM network as soon as it's acquired and matched to the booking by accession number. Drag-and-drop and ZIP import stay available for outside films.
Cases go to a radiologist, who reads in RadX's browser viewer — multi-pane layouts, MPR, MIP/MinIP, fusion, crosshair localisation, measurements, and keyboard shortcuts they can remap.
Candidate findings with confidence scores appear beside the images. The radiologist accepts or rejects each one; accepted findings compose into a draft in their own reporting style.
Structured report, templates, autosave, dictation. Sign-off stamps the radiologist's signature with date and time; a 7-day window allows corrections, then the report locks.
The signed report lives on the platform — send a secure link to the study and report to the referring doctor or the patient, print on the centre's letterhead, or export to Word.
Live status for every study, per-radiologist assignment, filters by patient, accession, modality and date — with scans arriving from the machine, a folder, or a ZIP.
Axial, coronal and sagittal reformats built in the browser; click a point and the crosshair localises the same anatomy in every plane. Slab MIP/MinIP, PET/CT-style fusion, and any slice captured straight into the report.
Accepted findings become a draft in your own words. Your templates, Radiopaedia and PubMed search beside the draft, Word export, print, and secure share links.
RadX is built around one rule: the radiologist decides. The AI does the legwork — the signature at the bottom is always human.
The AI reviews the images across series and planes and proposes candidate findings with its reasoning — a safety net against the miss at the end of a long shift.
Each finding is a card with a confidence score. Nothing enters the report without the radiologist's explicit approval.
Accepted findings compose into a full draft that follows the radiologist's phrasing and structure, learned from their signed reports.
Corrections and final reports quietly teach the AI the centre's standards — it becomes more useful the more you report.
RadX AI is made by practising radiologists and healthcare entrepreneurs — people who live the workflow this platform runs.
Over 10 years of reporting experience, specialising in whole-body and musculoskeletal (MSK) imaging.
Pharmaceutical entrepreneur with two start-ups under his belt, and CEO of SBF Pharma Pvt. Ltd.
10 years of reporting experience, with special interest in oncologic and breast imaging.
Diagnostic centres, hospital radiology departments and teleradiology desks — we'll set up a pilot on your own cases.
Tell us a little about yourself and we'll get back to you.
Your message is on its way to our inbox.
We usually reply within one working day.