Omi Speech

Medical speech-to-text, batch and live.

Transcribe consultations, dictation and calls with medical vocabulary, speaker labels and timestamps.

Try your audioRead the docs
Live transcriptionListening
Draft · Speaker A

start amoxicillin five hundred milligrams and repeat hba1c at review

Confirmed · 00:04

Start amoxicillin 500 mg and repeat HbA1c at review.

speaker labels onmedical vocabulary onrealtime
Flagship API

omi-medical-1

For production transcription where medical accuracy, availability and scaling are handled for you.

  • DeploymentHosted API or supported private cloud
  • IncludesBatch, realtime, vocabulary, speakers and timestamps
  • LanguagesEnglish benchmarked medically; seven more available for testing
#1 / 30medical benchmark
0.94%M-WER
97.7%dosage F1 · 86/89 events
Create an API keySee pricing

Omi and ElevenLabs are statistically tied on Medical WER. Omi recorded zero drug-name errors in 442 mentions.

Open model

omi-medical-edge-1

Quickly evaluate it through the hosted API, then run the same open English model inside your own environment.

  • DeploymentHosted API, Apple Silicon, NVIDIA CUDA or CPU
  • LicenceCC-BY-4.0 weights and MIT runtime for self-hosting
  • Hosted useEnglish batch transcription, word timestamps and optional speaker labels · $0.10/audio-hour after included usage
#2 openmedical benchmark
2.16%M-WER · served board
90.2%dosage F1 · 78/89 events
Try through the APIGet the model

The published served row is shown above. With unchanged weights, current local runtimes reach 6.54% WER on CUDA and 2.12% M-WER on MLX q8. Use the flagship for production features.