VideoShorts Transcriber
SkillMediaVideo transcription via faster-whisper. JSON + SRT with timecodes.
Available today. Use it from your connected AI after setup.
No other account needed.
Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.
Then ask your AI: use the VideoShorts Transcriber skill
What this skill tells your AI
The instructions your AI receives, as published by horosheff/hyperion-reels in skills/videoshorts-transcriber/SKILL.md and read by ahel’s review.
Вход
videoshorts-memory/00-brief.md— путь к видео, модель Whisper.cursor/videoshorts-handoff.md
Действия
- Прочитать
shared/agent-pipeline-pitfalls.md - Скопировать/проверить видео в
videoshorts-memory/input/если нужно - Запустить:
cd scripts
python transcribe.py "<video_path>" -o "../videoshorts-memory/transcripts/<stem>" -m <model>
- Проверить артефакты:
transcript.json— segments + опциональноwords[]/_words(karaoke ASS)transcript.srtaudio.wav
Env / language
VIDEOSHORTS_WHISPER_WORD_TIMESTAMPS=1(default) — word-level для субтитровVIDEOSHORTS_WHISPER_LANGUAGE=ru— только реальный ISO-код (ru,en, …)- Brief / UI
language: auto→ не передавать--language autoи не ставить env вauto/detect/none - Автоопределение =
language=Noneв faster-whisper (omit--language/ unset env) - Worker нормализует
auto/detect/пустую строку →None(не падает ValueError)
Ошибки
При падении Whisper — записать INC в pipeline-fix-queue.md, incident_report со ссылкой.
Signals
- GitHub stars
- 33
- Forks
- 25
- Last commit
- Sep 2026
Advanced
- Item type
- skill
- Key
videoshorts-transcriber- Source
- github.com/horosheff/hyperion-reels