Turn a video/audio file into speaker-labelled, word-accurate subtitles entirely offline. Three stages, each writing JSON into work/: audio ── transcribe.py ── work/transcript.json (faster-whisper, ...
角色 :内置涂鸦小男孩 Bo 与猫 Mochi,16 种表情,动作以"画出来 + 简单姿势"为主。 典型用途 :VideoScribe 风格的科普解说、教学、观念说明,配轻快旁白。 完整 API 速查见 whiteboard/SKILL.md。