Muse Voice Transcribe Language Lookup
Meta says Muse Voice Transcribe was trained with 70+ languages and extensively verified 25 for launch. Use this page as a lookup boundary: confirmed examples are not the same as a complete production support list.
Last verified · September 2, 2026
Indexed Meta developer documentation lists these 25 code-switching languages for Muse Voice Transcribe: Arabic, Bengali, Dutch, English, French, German, Hebrew, Hindi, Indonesian, Italian, Japanese, Kannada, Korean, Malay, Mandarin Chinese, Marathi, Polish, Portuguese, Spanish, Tagalog, Tamil, Telugu, Thai, Turkish, and Vietnamese.
- Chinese and English are shown together in Meta's code-switching example.
- Chinese, French, Hindi, Japanese, Spanish, and Vietnamese appear as public examples in Meta's launch page.
- Check the current logged-in Meta developer docs before committing to production coverage.
Training coverage describes the breadth of audio used to teach the model. It does not prove that every language has the same accuracy, evaluation depth, latency, punctuation behavior, or speaker-label quality.
Meta states that 25 languages were extensively verified at launch and recommends starting there. The list above should still be treated as a dated lookup because API availability, language-bias fields, and quality guidance can change.
Muse Voice Transcribe is documented as supporting arbitrary code-switching within a sentence or between sentences. This matters for multilingual meetings, dictation, support calls, and voice agents where speakers naturally mix languages.
Biasing can improve recognition of names, products, places, and domain-specific phrases. Confirm how the current API accepts those hints before building a workflow around them.