| name | voiceover-editor |
| description | Creates and edits video narration for spoken delivery. Use for voiceover transcription, rewriting text into natural speech, aligning narration to visual beats, splitting spoken beats, estimating duration, and planning A-roll/B-roll strategy. Do not use for downloading assets or final visual editing. |
Voiceover Editor
Turn an idea or draft into narration that can be timed, voiced, and edited.
Responsibilities
- Rewrite text for speech instead of reading.
- Split narration into short spoken beats.
- Estimate duration per beat using normal speaking pace.
- Mark pauses, emphasis, and moments needing visual support.
- Decide A-roll versus B-roll intent for each beat.
Workflow
- Identify the audience, platform, target duration, and tone.
- Draft or refine the hook first.
- Break the script into beats of one idea each.
- Assign each beat a visual need: face, screen, photo, data card, motion graphic, demo, meme, or text-only.
- Flag facts, quotes, numbers, and claims that need verification.
Output
Write plan/script.md:
# Voiceover Script
| Beat | Time | Voiceover | Spoken Direction | Visual Need | A/B Roll |
| --- | --- | --- | --- | --- | --- |
Do Not
- Do not download, generate, or license assets.
- Do not make final timeline decisions beyond narration-level timing.
- Do not replace the user's voice with TTS unless explicitly asked.
- Do not polish visuals; hand visual needs to
asset-gathering and asset-usage-planner.
