Can I transcribe telehealth video calls in real time?

Yes. Universal-3.5 Pro Realtime with Medical Mode transcribes during the visit, suitable for in-visit captions and ambient notes.

How does it handle poor telehealth video-call audio?

Universal-3.5 Pro is built for real-world audio, which matters when network compression degrades the signal.

Is it suitable for HIPAA-regulated telehealth?

AssemblyAI is a business associate under HIPAA and signs a standard BAA; PHI redaction across audio and transcripts; SOC 2 Type 2.

Does it support non-English telehealth visits?

Medical Mode supports English, Spanish, German, and French, including intra-utterance code-switching.

How much does real-time telehealth transcription cost?

$0.15/hr on top of the realtime or pre-recorded model — pay-as-you-go, no minimums.

How do I enable Medical Mode for telehealth transcription?

Add domain: "medical-v1" to your realtime session parameters.

Telehealth

Clinical-grade transcription for telehealth, live or recorded

Telehealth transcription turns live or recorded video-visit audio into an accurate clinical record. 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 — through accents, headsets, and compressed call audio. BAA available.

Video visits add variable audio, headsets, and network jitter on top of clinical vocabulary. Universal-3.5 Pro with Medical Mode holds a 3.2% Missed Entity Rate on that audio and runs in real time during the call — so your scribe, captions, or triage layer keeps up. One parameter, on async or realtime.

Try Medical Mode free on your audio Read the docs

Accuracy that survives real telehealth audio — accents, headsets, compression

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

Telehealth 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

Live captions and notes during the visit

Stream the call's audio into one WebSocket and get medical-grade transcripts back in real time — no model swap, no re-integration.

from assemblyai.streaming.v3 import StreamingClient, StreamingParameters

client = StreamingClient(api_key="YOUR_API_KEY")

# Telehealth: real-time transcription during the video visit
client.connect(StreamingParameters(
    speech_model="universal-3-5-pro",
    domain="medical-v1",
    keyterms_prompt=["telehealth", "hypertension", "lisinopril", "follow-up"],
))

Real-time, low latency

Streams on Universal-3.5 Pro Realtime for live captions and in-visit documentation.

Universal-3.5 Pro Realtime

Robust on messy audio

Holds accuracy through accents, headsets, and compressed call audio.

Universal-3.5 Pro

Multilingual visits

English, Spanish, German, French with native code-switching.

Medical Mode

Common questions

: Yes. Universal-3.5 Pro Realtime with Medical Mode transcribes during the visit, suitable for in-visit captions and ambient notes.
: Universal-3.5 Pro is built for real-world audio, which matters when network compression degrades the signal.
: AssemblyAI is a business associate under HIPAA and signs a standard BAA; PHI redaction across audio and transcripts; SOC 2 Type 2.
: Medical Mode supports English, Spanish, German, and French, including intra-utterance code-switching.
: $0.15/hr on top of the realtime or pre-recorded model — pay-as-you-go, no minimums.
: Add domain: "medical-v1" to your realtime session parameters.