Demo Narrate Reference

Shell Script Commands

• extract <video> [fps] -- extract frames and contact sheets
• tts-acts <dir> <timing.txt> [voice] -- generate per-act audio
• tts-acts --dry-run <dir> <timing.txt> -- show budgets without TTS
• merge-acts <video> <dir> <timing.txt> [out] -- merge audio onto video
• fade-intro <video> [secs] [out] -- add fade-in from black
• deps -- check dependencies
• voices -- list available TTS voices
• tts <file> [voice] -- single-file TTS (manual workflow, not used by pipeline)
• merge <video> <audio> [out] -- single-file merge (manual workflow, not used by pipeline)

Voice Options

Default voices for Step 4a preferences:

• en-US-AriaNeural -- warm female (default)
• en-US-GuyNeural -- male
• en-US-JennyNeural -- friendly female
• en-GB-SoniaNeural -- British

For the full list of available voices, run "$NARRATE_SH" voices.

Output Directory Structure

<output_dir>/
  act1_opening.txt          <- plain text, spoken words only
  act1_opening.mp3          <- TTS audio for act 1
  act2_extraction.txt
  act2_extraction.mp3
  ...
  timing.txt                <- "filename offset_seconds" per line
  script_final.txt          <- master reference with all acts + metadata

Standalone Installation

This skill can also be used without the full plugin:

1. Copy SKILL.md to ~/.claude/commands/demo-narrate.md
2. Copy scripts/demo-narrate.sh to ~/.local/bin/demo-narrate.sh and chmod +x it (must be on your PATH)
3. The fallback in the "Shell Script" section will find it via command -v demo-narrate.sh -- no path edits needed