Key Info

Mistral used its Voxtral 3 Transcribe model to capture every main-stage talk in real time at AI Engineer Paris, and says the same technology is available for other events and apps.

Highlights

  • Voxtral 3 Transcribe handled live, real-time transcription of all main-stage talks at AI Engineer Paris.
  • Accuracy was improved via context biasing: feeding the model speaker names and technical terms drawn from the conference agenda.
  • Mistral notes the same transcription technology can be used for users' own events and applications.