Dictation Software Built for
Radiology Reports
Read the image. Dictate the report. VoicePrivate transcribes it locally, with no cloud transcription server and no internet required.
Built for radiology workflows
Whether you're reading 20 studies a day or 80, the documentation burden is the same. VoicePrivate cuts the time between reading and report without touching your network.
Who uses it
- Hospital and academic radiologists dictating structured reports
- Teleradiologists working across multiple health systems
- Private practice radiologists replacing enterprise dictation systems
- Radiology residents building their documentation workflow
- Radiology practice administrators managing dictation costs
The workflow
- Open the study in your PACS viewer
- Press your hotkey, then dictate your findings and impression
- VoicePrivate transcribes locally, under 2 seconds
- Text pastes directly into your report interface or EHR
- No audio file created, no data transmitted, no log on any server
Radiology reports require careful data handling
Patient ID, study type, imaging findings, impressions, and recommendations can include protected health information. Cloud dictation sends that audio and text through a third-party server.
Cloud dictation in radiology
- Patient audio uploaded to vendor servers
- Clinical content processed outside your local device
- Vendor security review and data-processing terms to evaluate
- Data retention liability on third-party infrastructure
- Fails in restricted hospital networks with no internet access
VoicePrivate - on-device processing
- All audio processed locally on your Mac
- Dictated audio and generated text stay local to your device
- Designed to minimize third-party handling of dictated content
- Operates without internet in secure hospital environments
- Tamper-evident audit log proves every transcription stayed local
This is structural product behavior, not a compliance certification. The audio physically does not leave your device. You can verify it: run a network monitor during transcription and observe zero outbound traffic. No cloud dictation tool can make this claim.
Fits your reading room, not the other way around
No cloud transcription server. No enterprise onboarding for individual use. A simpler architecture to review with your privacy officer.
Mac-native
Built for Mac, including M-series chips. Runs as a lightweight background process. Activate with a global hotkey from inside any application, your PACS viewer, web-based RIS, Epic, or a plain text editor.
Offline by design
No internet required. The speech model runs entirely on your device. Works in restricted hospital environments where external network access is blocked or limited. Same performance on and off VPN.
Radiology vocabulary
The healthcare dictionary includes radiology-specific terminology: imaging modalities (CT, MRI, PET, fluoroscopy, ultrasound), anatomical structures, contrast agents, measurement language, and impression formatting. Accurate out of the box, extensible with a custom dictionary.
Accuracy that holds up in clinical language
The healthcare dictionary covers radiology at the level you need: modality-specific terms (T2-weighted, FLAIR, diffusion-weighted imaging), contrast phases (arterial, portal venous, delayed), anatomical landmarks, measurement conventions, and impression language.
Grammar polish cleans up natural speech patterns, including false starts, filler words, and dictation-style phrasing, so the transcription reads like a finalized report rather than a raw transcript.
The custom dictionary extends coverage further. Add abbreviations, institutional terms, referring physician names, or subspecialty vocabulary your reading group uses consistently.
How it compares
Most radiologists are either on Dragon Medical One or evaluating whether they can justify the enterprise cost. Here is the full picture.
| Feature | VoicePrivate Healthcare | Dragon Medical One | Suki AI | Otter.ai Health |
|---|---|---|---|---|
| On-device processing | ✓ 100% | ✗ Cloud | ✗ Cloud | ✗ Cloud |
| Vendor cloud processing | ✓ No cloud transcription server | ✗ Yes | ✗ Yes | ✗ Yes |
| Vendor review surface | ✓ Local dictation path | ✗ Cloud service | ✗ Cloud service | ✗ Cloud service |
| Works offline | ✓ Yes | ✗ No | ✗ No | ✗ No |
| Mac support | ✓ Full (M-series) | ✓ Yes | ✓ Yes | ✓ Yes |
| Radiology vocabulary | ✓ Included | ✓ Included | ✓ Included | Partial |
| Custom dictionary | ✓ | ✓ | ✗ | ✗ |
| Works without IT procurement | ✓ Download and use | ✗ Enterprise contract | ✗ Enterprise contract | ✗ Enterprise contract |
| Annual price per seat | $297/yr | $600–900+/yr | Enterprise pricing | Enterprise pricing |
Cloud dependency and vendor architecture based on published documentation. Verify before making procurement decisions.
Straightforward pricing
No enterprise contract. No procurement cycle. Download and try it free, then decide.
Compare: Dragon Medical One runs $600–900+/yr per seat, requires enterprise IT involvement, and processes audio in the cloud.
Questions from radiologists
How should my team evaluate this for compliance?
VoicePrivate processes dictation locally on your device rather than sending audio to a transcription server. Compliance requirements depend on your workflow, policies, and deployment environment, so your compliance team should evaluate the product for your use case. This is product architecture information, not a certification.
Will it work in our hospital's restricted network?
Yes. VoicePrivate requires no internet connection to transcribe. The speech model runs entirely on your device. It will function normally in environments where external network access is blocked or filtered, including secure reading room networks.
Does it know radiology terminology?
Yes. The healthcare dictionary includes radiology-specific vocabulary: CT, MRI, PET, fluoroscopy, ultrasound, anatomical structures, contrast agents, measurement terms, and impression language. For subspecialty or institutional terms not in the dictionary, you can add them via the custom dictionary.
What microphones work with it?
Any microphone your Mac recognizes: internal microphone, USB headsets, dictation microphones, or external mics connected via USB or Bluetooth. VoicePrivate uses the macOS system audio input, so any mic that works on your Mac will work with VoicePrivate.
Can it paste text directly into my PACS or RIS?
Yes. VoicePrivate uses system-level keyboard simulation to type transcribed text into whatever application is active, including PACS viewers, web-based RIS interfaces, Epic, and plain text editors. There is no PACS-specific integration required.