Skip to main content
Set a primary language for best accuracy, or use Auto when you switch languages between recordings. Language support depends on your selected voice model. A language offered by a cloud provider is not necessarily supported by every offline model. Transcription language controls what you speak; a translation mode controls the language of the final written output.

​
Language Examples

Utter’s available models cover languages including those below. Check the language choices for your selected model before a long recording; this list is not a promise that each model supports every language, dialect, or mixed-language recording.
  • English (US, UK, AU)
  • Spanish
  • French
  • German
  • Italian
  • Portuguese
  • Japanese
  • Korean
  • Chinese (Mandarin)
  • Hindi
  • Arabic

​
Language Settings

​
Automatic Detection

If you set your language to Auto in settings, Utter will automatically detect the language you’re speaking. This works well when:
  • You consistently speak one language
  • You switch languages between recordings

​
Setting a Primary Language

For better accuracy, set your primary language:
1

Open Settings

Click Utter in the menu bar and choose Settings.
2

Go to Language

On Mac, open Settings > General > Language, then Transcription language. On iPhone, edit the mode and open Advanced Settings > Voice Language.
3

Select Language

Choose your spoken language. If the mode has its own language setting, check that setting too rather than relying only on the default.
What to expect: New recordings are transcribed using your selected language.
Test one short recording after changing models or languages. Automatic detection can struggle with very short phrases, names, or several languages in the same recording.

​
Translation

Use a custom mode with a compatible AI model to translate the transcribed text. The voice model must first understand your spoken language; the AI model must support the requested output language.

​
How to Set Up Translation

1

Create a Custom Mode

Go to the main screen and create a new mode.
2

Name Your Mode

Give it a clear name like “Translate to Spanish”.
3

Set Mode Instructions

In the mode instructions, write: Translate to Spanish (or your target language).
4

Use the Mode

Select the mode before dictating.
What to expect: The output text is translated to the target language you specified.
You can speak in any supported language. Utter transcribes first, then applies the mode instructions to translate.

​
If the Language or Output Is Wrong

  1. Check the original transcript before the translated output. If the original is wrong, change the voice language or model first.
  2. If the original is correct but the translation is wrong, check the mode instructions and AI model. For example: Translate the transcript into Spanish. Return only the translation.
  3. For offline use, confirm both the voice model and AI model support your languages and have finished downloading. Follow the offline setup test.
  4. Add repeatedly misspelled names to custom vocabulary. Review translations of names and technical terms before sharing them.

​
Language-Specific Tips

Works best with clear pronunciation. Supports US, UK, and Australian accents.
Select a voice model that supports the variety you speak. If you need Simplified or Traditional Chinese output, state that explicitly in your mode instructions and check the result.
Automatically handles Kanji, Hiragana, and Katakana output.
Supports Modern Standard Arabic and many regional dialects.

​
Still Need Help?

Contact support with your spoken language, desired output language, selected voice and AI models, and a short non-sensitive example.