On the roadmap, not built yet.
Transcription already works in other languages: Whisper auto-detects, or you pick a language in Settings (it defaults to English because auto-detect occasionally mislabels English as a similar-sounding language). Captions then burn in whatever Whisper returned.
What's missing is translation: taking an English video and burning in Spanish captions, or vice versa.
What's involved
- Translate the transcript while keeping word-level timings usable. This is the real problem: word counts and word order change between languages, so per-word karaoke timing cannot be carried across directly. Translating per caption group and redistributing timing across the group proportionally is probably the pragmatic answer.
- A target-language picker, per clip rather than per project, since one source video may want several language versions.
- Font coverage. The bundled caption fonts do not all cover non-Latin scripts, so CJK, Arabic or Cyrillic output needs a fallback. Note that Arabic and Hebrew also need right-to-left handling in libass.
Start at src/shared/captionLayout.ts for grouping and src/main/pipeline/captions.ts for ASS generation.
On the roadmap, not built yet.
Transcription already works in other languages: Whisper auto-detects, or you pick a language in Settings (it defaults to English because auto-detect occasionally mislabels English as a similar-sounding language). Captions then burn in whatever Whisper returned.
What's missing is translation: taking an English video and burning in Spanish captions, or vice versa.
What's involved
Start at
src/shared/captionLayout.tsfor grouping andsrc/main/pipeline/captions.tsfor ASS generation.