screenkite-audio-dub

Replace a ScreenKite recording's spoken audio with a time-matched dub in another language. Original pauses stay. Lip-ish alignment is per cue, not a single stretched file.

1. export-words (or Scribe)     → word-level JSON
2. pack_cues.py                 → cues.json + source.srt
3. fill dub_text (agent/human)  → translated lines, same timestamps
4. gen_dub.py                   → mix.wav (TTS + speed fit + silence bed)
5. dry-run table                → confirm before mutating
6. apply_dub.py                 → mute mic + add language-tagged track

Prerequisites