Back to the index

Voice dictation for reports

CBCTHub includes speech recognition built into the report editor. You speak, and the text appears. It has an embedded dental glossary so words like “periapical”, “mandibular”, “inferior alveolar” transcribe correctly.

NotaRecognition runs 100% in your browser (local Whisper model) — the audio never leaves your device. It does not require connection to external services such as Google Speech or AWS.

Enable dictation

  1. 1
    In the viewer's report editor, look for the microphone icon in the editor bar.
  2. 2
    The first time, the browser asks for permission to use the microphone. Accept.
  3. 3
    Also the first time, the model downloads in the background (around 150 MB). After that it is cached and starts in seconds.
  4. 4
    Click the microphone to start. Speak clearly at normal speed — the text appears at the cursor as you speak.
  5. 5
    Click the microphone again to stop.

Dental glossary

CBCTHub improves recognition with a dentistry-specific glossary. Common terms are recognized correctly without spelling them out:

periapical, periodontal, inferior alveolar canal, mental foramen, maxillary sinus, temporomandibular joint, mesial root, distal root, furcation, alveolar crest, cortical bone, cancellous bone, radiolucent lesion, radiopaque lesion, etc.

TipYou can add your own terms to the glossary from Settings › Radiology report — useful if you use regional terminology or commercial names of specific implants/systems.

Limitations

  • Works best in neutral Spanish/English; very regional accents may need more corrections.
  • You need a decent microphone — the laptop one is usually enough; a headset is better.
  • The first model load is slow. After that it is instant.