● AI clinical notes — radiology · dental · medical

Stop typing notes. Start seeing patients.

Radnx is the AI medical scribe for radiology, dental and medical teams. Speak the way you always have — and get a clean, structured, coded note back in seconds, in 60+ languages, running on infrastructure you control.

Explore apps

Less time charting. Faster turnaround. Notes that read like you wrote them.

Self-hostable 60+ languages No PHI stored server-side
Radnx — clinical note

Dictation

"No acute findings. Small simple cyst right kidney. Impression: unremarkable study…"

Structured note

Coding

ICD-10Template

One scribe, every clinical setting

Radiology Dental Medical Hospitals & clinics Private practice
Built for your specialty

One scribe. Every specialty.

The same dictate-to-note engine, tuned for how radiologists, dentists and clinicians each document. All three are live today.

Live

Radiology

Multilingual & French-tuned
  • RSNA-structured reports + ICD-10
  • 60+ languages; fine-tuned French model
  • Templates, correction & chat assistant
Live

Dental

Dental AI notes
  • Exam, treatment-plan & SOAP notes from dictation
  • Charting-friendly, structured output
  • Works at the chairside on any device
Live

Medical

Medical notes
  • SOAP & progress notes for clinics
  • Specialty templates & coding
  • Multilingual encounters
Why Radnx

From spoken findings to a sign-off-ready note

Everything between the clinician's voice and a finished note — handled, accurately, in the language they think in.

Dictate, don't type

Speak naturally. Speech-to-text tuned for clinical vocabulary captures anatomy, laterality, measurements and units exactly as spoken.

Structured & coded, automatically

Notes come back in the right format — RSNA for radiology, SOAP and specialty templates for dental and medical — with suggested ICD-10 codes. No reformatting, every time.

Truly multilingual

60+ languages out of the box, plus a dedicated French radiology model fine-tuned on domain audio for accuracy generalists can't match.

Private by design

Run it on your own servers or on-prem. Audio and notes aren't persisted server-side, and logs redact protected health information.

Templates + an AI co-editor

Auto-detect the right template, fix dictation slips automatically, and refine wording with a chat assistant that edits the note inline as you watch.

Everywhere you work

Web, desktop, Android, iOS and a browser extension — the same workflow at the workstation, chairside, or on the ward.

How it works

Three steps. Seconds, not minutes.

STEP 01

Dictate the encounter

Hit record and describe the study or visit as you normally would — from the web app, desktop, phone, or the browser extension.

STEP 02

We transcribe & correct

Clinical-tuned speech-to-text transcribes the audio — self-hosted for French — then a correction pass fixes the terms generic models get wrong.

STEP 03

Get a structured note

An LLM assembles the note in your specialty's format with coding. Review, refine with the chat assistant, and sign off.

3
Specialties: radiology · dental · medical
60+
Languages supported
5
Platforms: web, desktop, Android, iOS, extension
100%
Self-hostable on your infrastructure
For hospitals & teams

Deploy department-wide, on your terms

Radnx was built to run inside your network, not just in someone else's cloud. Keep data where compliance requires it and scale across radiology, dental and medical teams.

  • On-prem or private-cloud deployment (Docker + nginx)
  • Bring-your-own model: serve the reporting LLM where you choose
  • PHI-safe logging & per-IP rate limiting built in
  • Rollout support for desktop, mobile and the browser extension
Deployment topology

Your network

Web + API behind your nginx / TLS

Speech-to-text

Self-hosted on your GPU

Reporting model

LLM endpoint you control

Voices

Less admin. More medicine.

"The win isn't dictation — it's that the structure and coding are already done when I stop talking."

AR
Consultant Radiologist
Illustrative

"Chairside notes that write themselves mean I finish the day without an hour of charting."

DM
Dentist
Illustrative

"Running it inside our own network was the only way this passed review. Radnx made that the default."

IT
Clinical IT Director
Illustrative
Pricing

Simple plans, built around your deployment

Final pricing is set with you based on specialty, volume and hosting. Get in touch to get started.

Solo

Let's talk / clinician

For individual clinicians and small practices.

  • Any one specialty (web)
  • Structured notes + ICD-10
  • Desktop & browser extension
  • Email support
Get started

Clinic

Custom / team

For clinics, imaging centers and groups.

  • Everything in Solo
  • Multiple specialties & shared templates
  • Mobile & iOS apps
  • Priority support & onboarding
Get started

Enterprise

Custom

For hospitals and networks with compliance needs.

  • On-prem / private-cloud deployment
  • Bring-your-own model & GPU
  • SSO & security review support
  • Dedicated rollout & SLA
Talk to sales
FAQ

Questions, answered

Which specialties does Radnx support? +
All three are live: radiology (multilingual + a French-tuned model), dental, and medical notes. Open them from the Specialties section above — just pick your specialty.
Is patient data sent to a third-party cloud? +
You decide. Radnx can run entirely inside your own network — the web app, API, speech-to-text, and the reporting model can all be self-hosted. Audio and notes aren't persisted server-side, and logs redact protected health information.
How accurate is it for non-English documentation? +
60+ languages with clinical vocabulary biasing, plus a speech model fine-tuned on French radiology audio and a correction pass for domain terms generic models mishear.
Does it produce structured notes or just a transcript? +
Both. You get the verbatim transcript and a generated note in your specialty's format (RSNA, SOAP, or templates) with suggested ICD-10 codes — refinable with the chat assistant.
Which devices are supported? +
Web, desktop (Windows/macOS/Linux), Android, iOS, and a browser extension that works alongside your PACS, RIS or practice software.
Can we use our own LLM or GPU? +
Yes. The reporting model talks to any OpenAI-compatible endpoint and the speech-to-text runs on your GPU — all environment-configured, no code changes.
Get in touch

See Radnx on your own cases

Tell us your specialty and we'll tailor a walkthrough — or jump straight into the live apps and try it on a real dictation today.

Prefer email? hello@radnx.ai