New Voice Standard Bench: 50+ LLMs ranked on real voice tasks Learn more
Dental

Documentation that speaks dental

Dental transcription converts chairside and operatory speech into accurate chart notes, including tooth numbers, surfaces, and procedure codes. AssemblyAI's Medical Mode, a one-parameter add-on to Universal-3.5 Pro, holds a 3.2% Missed Entity Rate on clinical conversations — the lowest across benchmarked providers. BAA available.

Tooth numbering, procedure codes, and dental terminology trip up general transcription. Universal-3.5 Pro with Medical Mode captures the clinical vocabulary of the operatory, so chart notes and treatment plans start accurate. One parameter, async or chairside realtime.

The same medical entity accuracy, applied to dental terminology

Missed Entity Rate 3.2%

Missed Entity Rate on clinical conversations with Medical Mode.

Fewer missed entities ~20%

Fewer missed medical entities than the base model without Medical Mode.

Fewer entity errors 87%

Fewer medical entity errors than the base model.

Benchmarked providers #1

Lowest Missed Entity Rate vs. Deepgram, Speechmatics, AWS, and Google.

Benchmarks

Dental transcription accuracy vs. other providers

Missed Entity Rate is the share of clinical entities — medication names, dosages, diagnoses — that a model fails to transcribe correctly. Lower is better.

Missed Entity Rate and Word Error Rate on clinical conversation audio Lower is better. Medical Mode enabled via domain: "medical-v1".
Provider Missed Entity Rate Word Error Rate
AssemblyAI Universal-3.5 Pro w/ Medical Mode 3.2% 5.3%
Speechmatics Enhanced Medical 3.6% 5.5%
Deepgram Nova-3 Medical 4.7% 6.1%
AWS Transcribe Medical 8.7% 5.9%
Google Medical Conversation 24.4% 12.9%
See the full benchmarks
Activation

Tune to dental vocabulary

Pass operatory vocabulary as keyterms — tooth numbers, surfaces, procedures — one parameter, no model swap.

import assemblyai as aai

aai.settings.api_key = "YOUR_API_KEY"

# Dental: tooth numbering, procedures, materials
config = aai.TranscriptionConfig(
    speech_models=["universal-3-5-pro", "universal-2"],
    domain="medical-v1",
    keyterms_prompt=["distal", "occlusal", "composite restoration", "#14"],
)

transcript = aai.Transcriber().transcribe("./dental-exam.wav", config)
print(transcript.text)
Worked example

Tooth numbers and CDT codes that survive the transcript

General transcription renders number-fourteen-M-O as prose and drops the procedure code entirely. Medical Mode plus keyterms holds the tooth numbers, surfaces, and CDT codes that make an operatory note billable.

Chairside dictation -> operatory note

#14 MO composite, D2392. Caries to dentin on the
mesio-occlusal surface. Local: 1 carpule 2% lidocaine
w/ 1:100,000 epi. Rubber dam isolation.

#19 DO amalgam noted, margins intact, no treatment.
#30 recommend crown, D2740 - patient to schedule.

Perio: generalized 3-4mm probing depths, localized
5mm on #2 and #15 distal. Recall 4 months.

What has to survive transcription: tooth numbers
(#14, #19, #30, #2, #15), CDT codes (D2392, D2740),
surfaces (MO, DO, distal), and probing depths.

Dental terminology

Surfaces, tooth numbers, and procedure terms captured for accurate chart notes.

Universal-3.5 Pro

Chairside or recorded

Stream during the exam or transcribe recordings; robust on operatory audio.

Universal-3.6 Pro Realtime

Private by default

PHI redaction, BAA available, SOC 2 Type 2.

Medical Mode

Trusted by the world’s leading enterprises

Metaview
Dovetail
Granola
Apollo.io
Ashby
Siro
Calabrio
Cluely
Genio
Commure
Retell
CallRail
LiveKit
EliseAI
ClickUp
HeyGen
Metaview
Dovetail
Granola
Apollo.io
Ashby
Siro
Calabrio
Cluely
Genio
Commure
Retell
CallRail
LiveKit
EliseAI
ClickUp
HeyGen
Metaview
Dovetail
Granola
Apollo.io
Ashby
Siro
Calabrio
Cluely
Genio
Commure
Retell
CallRail
LiveKit
EliseAI
ClickUp
HeyGen
Metaview
Dovetail
Granola
Apollo.io
Ashby
Siro
Calabrio
Cluely
Genio
Commure
Retell
CallRail
LiveKit
EliseAI
ClickUp
HeyGen

Common questions