Live
Medical Mode
A correction pass for clinical language that fixes the drug names, procedures, and conditions the model heard but misspelled.
$0.15 /hr
added to your speech-to-text model
Healthcare AssemblyAI for
AssemblyAI for
Healthcare
Clinical-grade voice AI for scribes, dictation, and voice agents.
“We've integrated the newest models from AssemblyAI for pre-recorded audio ASR in our ambient product, and it's been excellent.”
75%
engineering time savings on infrastructure
“We've integrated the newest models from AssemblyAI for pre-recorded audio ASR in our ambient product, and it's been excellent.”
75%
engineering time savings on infrastructure
“We've integrated the newest models from AssemblyAI for pre-recorded audio ASR in our ambient product, and it's been excellent.”
75%
engineering time savings on infrastructure
“We've integrated the newest models from AssemblyAI for pre-recorded audio ASR in our ambient product, and it's been excellent.”
75%
engineering time savings on infrastructure
Use cases
Clinical accuracy for anything you build
Everything your clinical product needs to listen and interact from a single platform.
Healthcare voice agents
Use realtime with Medical Mode, so scheduling and intake agents catch medication and condition names correctly.
Learn moreAmbient AI scribes
Transcribe the visit with Medical Mode and speaker labels, then draft the note with LLM Gateway.
Learn moreClinical transcription
Run Medical Mode on recorded audio for records, research, and review, with speaker separation for multi-party encounters.
Learn moreMedical dictation
Use the Dictation API for fast, cleaned-up dictation in your app, formatted by the instructions you send.
Learn moreVeterinary scribes
Vet Mode corrects vet drug names that general models confuse with human ones, so the chart names the right drug.
Learn more Patient data
Privacy your patients expect
Enterprise-ready infrastructure with the protections your compliance team asks about first.
BAA available
AssemblyAI is a business associate under HIPAA and signs a standard BAA for customers processing PHI.
PHI redaction
Redact PHI across audio and transcripts before the data reaches your downstream systems.
SOC 2 Type 2
Independently audited security controls, with encryption in transit and at rest and configurable data retention.
Data control
Process patient audio with zero data retention, or self-host the models so data stays in your network.
Benchmarks
See the full benchmarks First on clinical accuracy
3.2% Missed Entity Rate on medications, dosages, and diagnoses. Lower is better.
| AssemblyAI Universal-3.5 Pro w/ Medical Mode | Speechmatics Enhanced Medical | Deepgram Nova-3 Medical | AWS Transcribe Medical | Google Medical Conversation | |
|---|---|---|---|---|---|
| MER | 3.2% | 3.6% | 4.7% | 8.7% | 24.4% |
| WER | 5.3% | 5.5% | 6.1% | 5.9% | 12.9% |
Modes
Go deeper where accuracy matters most
Get the highest accuracy transcription for your patients with one additional field.
Beta
Vet Mode
The same correction pass for veterinary terms, including the vet drugs a general model swaps for similar-sounding human ones.
$0.15 /hr
added to your speech-to-text model
Docs
Ship it in an afternoon
Simply set the domain on a pre-recorded request or as a streaming connection parameter.
Domain 
import requests
response = requests.post(
"https://api.assemblyai.com/v2/transcript",
headers={"authorization": "YOUR_API_KEY"},
json={
"audio_url": "https://example.com/clinical-visit.mp3",
"speech_models": ["universal-3-5-pro"],
"domain": "medical-v1", # or "veterinary-v1"
},
)
print(response.json()["id"])
Playground Try it out
Hear the difference on your own audio
Run your clinical audio through Medical Mode in the no-code playground.
