Meta Launches Real-Time Muse Voice Transcribe API at $3 Per 1,000 Minutes

Meta introduced Muse Voice Transcribe, a real-time audio perception model capable of natively handling streaming speech-to-text, speaker diarization, and endpointing in a single system. The model is currently available to developers through the Meta Model API, priced at $3 for 1,000 audio minutes. The system can isolate and transcribe audio for more than 20 concurrent speakers during hour-long sessions. It is trained across more than 70 languages, with 25 explicitly validated at launch, and can seamlessly track mid-sentence code-switching between different dialects. "The model decides when to listen. It waits a little longer on hard words and commits faster on easy ones, using adaptive delay to predict each token and increase accuracy." — Mark Zuckerberg