Languages and translation

TranscribeX needs to know which language is spoken in your audio. After transcribing, you can translate the transcript into another language (Pro) and show the original, the translation, or both.

Set the transcription language

The language list depends on the model, because each model supports different languages.

• Default for new jobs: Settings › Transcription › Default Target Language. Despite the name, this is the language spoken in the audio, not a translation target. The list matches your Default Model.
• For one file: open the transcript and use the language menu next to the model menu at the top. Then click Start, or Restart to transcribe again.

If you switch to a model that doesn’t support the selected language, TranscribeX picks the first language that model supports. Check the language before you start.

Auto-detect the language

Choose Auto Detect at the top of the language menu to let the model work out the language. It’s available on multilingual models, including the multilingual Whisper models, Parakeet V3, the Qwen3 models, the Mistral and ElevenLabs online models, and the Nemotron streaming model. English-only models (the Whisper EN models and Parakeet V2), the distilled and language-tuned Whisper models, Cohere and Apple On-Device don’t offer it.

Auto-detect works best when one language is spoken clearly from the start. If a recording starts with silence or music, or mixes languages, choose the language yourself.

Translate a transcript (Pro)

1. Open a finished transcript.
2. In the right panel, click Overview, then the Translate tab.
3. Pick an Engine: Apple, DeepL or AI. With AI, also choose an AI Instance:.
4. Pick the target Language:.
5. Click Translate. A progress bar shows the percentage, and Cancel stops it.

Translate is disabled when the target language is the same as the transcript’s language. The Settings link next to the engine opens that engine’s settings. Without Pro, clicking Translate opens the Pro upgrade sheet.

Note: If you leave the transcript while a translation is running, TranscribeX asks whether to leave anyway. Leaving cancels the translation.

Choose a translation engine

Set the default engine for new transcripts in Settings › Translation › Translation Engine. You can still switch engines per transcript in the Translate tab.

Apple

Apple Translation runs on your Mac and needs macOS 15 Sequoia or later. It uses language packages that macOS downloads once. If a package is missing when you click Translate, macOS asks to download it. Let the download finish, then click Translate again.

To download packages ahead of time:

1. Open Settings › Translation and choose Apple.
2. Under Apple Language Packages, click Download next to each language you need. Installed languages show Installed.
3. Click Refresh if a status looks out of date.

To remove a package later, go to System Settings › General › Language & Region › Translation Languages.

DeepL

DeepL needs your own DeepL API key. In the Translation pane, choose DeepL, paste your key into API Key, and set Plan to Free or Paid to match your DeepL account. Get DeepL API Key opens the DeepL API page. Your transcript text is sent to DeepL to translate.

AI

The AI engine uses the AI services you set up in Settings › AI Service, including the built-in Local LLM. If none is set up, the Translate tab says “No AI instances configured.” See AI services.

Show the original, the translation or both

In the Translate tab, the Show: switch turns the translation on or off in the transcript view. It turns on automatically when an Apple translation starts.

• Segments view: the translation appears in gray under each paragraph. With Allow Edit on, you can click it and correct it.
• Text view: open Overview › Settings and use the Language picker in the Text section to switch between Original and Translated.
• Exports: choose Original Language, Translated Language or Both Languages in the export window. See exporting transcripts.

Live translation

You can also translate captions while you record (Pro). That’s set up in the live transcription view’s Caption Settings. See live transcription.