ARCDUB · VIDEO DUBBING
Speak another language in your original voice
Transcription, translation, voice-cloned dubbing, subtitles, and lip sync—use any module standalone or run the full pipeline end to end.
Product walkthrough video
PIPELINE
Speech recognition
Sentence-level timestamps and speaker diarization
Translation
Preserves domain terms while matching original pacing
Voice-cloned dubbing
Each speaker retains their original vocal timbre
Lip sync
Syncs lip movements naturally to the new audio
Mix & export
Audio mixing, burned-in subtitles, and final video
Three standalone modules
Dubbing
Every speaker keeps their own voice, with pacing automatically aligned to the original clip.
Subtitles
Bilingual subtitle workbench—proofread line by line, then export SRT or burn subtitles in.
Lip sync
Matches lip movements naturally to the dubbed audio—built for talking-head and interview videos.
BENCHMARKS
Supported languages
3 languages
Chinese · English · Thai
Speech recognition
2.2% WER
Lower is better · FLEURS
Voice naturalness
4.4 MOS
Cross-lingual · 5-point scale
Processing time (10-min video)
15 min
End-to-end · ASR to final export