Otter now supports transcription in languages including Mandarin Chinese, Korean, German, Portuguese and Turkish. Arabic, Hindi, Thai and Persian are among the languages due to arrive later.
The new languages run on Otter’s own speech models, trained using publicly available datasets. Native-speaking volunteers also helped the company assess transcript accuracy.
The planned multilingual feature will detect language changes automatically, without users adjusting settings during a meeting. For companies whose teams switch languages in conversation, that could make meeting records easier to maintain, but the capability is not yet available.