Meta launches Muse Voice Transcribe with support for 5 Indian languages

On Tuesday, September 1, Meta, a leading tech company, unveiled Muse Voice Transcribe, an advanced speech-to-text model designed to transcribe conversations in real-time. This innovative model, developed by Meta Superintelligence Labs (MSL), boasts the ability to handle conversations across multiple languages simultaneously, making it an invaluable tool for multilingual markets like India.

Muse Voice Transcribe supports over 70 languages, including Hindi, Tamil, Telugu, Malayalam, and Kannada, five major Indian languages. The model is adept at understanding code-switching, where speakers switch between languages within a single conversation, all without requiring separate models or additional processing. This feature is especially beneficial in India, where English and regional languages coexist.

One of the standout features of Muse Voice Transcribe is its capacity to transcribe speech as it is spoken, a process known as streaming transcription. This means text is generated in real-time, eliminating the need to wait for the entire recording to be processed. Additionally, the model can distinguish between multiple speakers in recordings involving more than 20 voices, and it can handle recordings longer than an hour.

Muse Voice Transcribe stands out for its balance between speed and accuracy. Unlike models that use a fixed setting for processing time, Muse Voice Transcribe dynamically adjusts its listening time based on each word, allowing it to swiftly respond to straightforward speech while taking more time for more challenging sounds.

The model has been trained across over 70 languages, with 25 languages validated at launch. As of September 1, 2026, Muse Voice Transcribe ranked first on the Artificial Analysis streaming speech-to-text leaderboard. Users can access the model through Meta's Model API, available starting from the launch date. The API, priced at $3 per 1,000 audio minutes, translates to approximately $0.18 per hour.

This launch of Muse Voice Transcribe is particularly timely, as speech-based AI systems gain significance in various applications, including transcription, dictation, coding, and voice assistants. The ability to support code-switching and a broader range of regional languages can enhance the comprehension of multilingual communication, making such systems more effective in diverse markets like India.

Furthermore, Meta has provided additional technical insights about Muse Voice Transcribe in a research blog, accompanying the launch.