Languages and automatic detection

Choose the language you speak

Whisper One 0.1.12 supports Chinese, English, German, Portuguese, French and Hindi. Both Free and Pro include these languages. Your plan changes the usage allowance, not the available recognition languages.

New installations use Auto detect. To choose a language yourself, open Settings → Language → Default transcription language. Your choice is saved for the next recording. Updating an existing installation keeps its saved preference.

French and Hindi

The Hindi option is currently labeled “preview” in the app. For French or Hindi dictation, select that language in Settings to use the configuration tested for it. Automatic detection and manual language selection can give different results. Try a short message in your usual speaking style before recording longer material.

Chinese-English mixed speech

Use Auto detect when Chinese sentences include English names, product terms or technical vocabulary.

明天下午 review 这个 proposal，确认 budget 之后再发给 Alex。

Read the transcript before sending. A short phrase, unfamiliar name, accent or background noise can change the result. Hindi-English mixed speech is not included in the current release; Hindi support does not imply support for every Indian language or language combination.

Recognition, interface and translation

The recognition setting describes what you say. The app interface can be English or Simplified Chinese independently of that setting. Ordinary dictation writes what you say; choosing another recognition language is not a translation command. When asking an AI assistant for a translated draft, include that request in your prompt.

If a result uses the wrong language, select the intended language, check the microphone and try a short recording again. Keep names and numbers in mind when reviewing the result. Smart polishing is optional and starts off by default; it does not replace proofreading.

See the changelog for version-specific changes, or troubleshooting for recording and insertion issues.