YouTube Voiceover
Inputs
Use script.md, target language, preferred voice, speed, emotion, and any tool-specific TTS requirements.
Process
- Extract narration only from the script.
- Clean text for speech: remove table syntax, visual notes, and unsupported symbols.
- Add pronunciation notes for names, acronyms, numbers, and uncommon terms.
- Split narration into timed segments.
- Generate audio if a TTS tool is available; otherwise create a TTS-ready plan.
- Record file names and durations.
Output Format
Create voiceover-plan.md:
# Voiceover Plan
## Voice Direction
## Clean Narration
## Pronunciation Notes
## Segments
| Segment | Text | Target Duration | Audio File |
|---|---|---:|---|
## Audio Settings
## Checks
Quality Check
- The narration must be speakable without markup.
- Segment durations should support the target video length.
- Voice direction should match the audience and topic.
- Any generated audio should be checked for clipping, truncation, and mispronunciation.