---
name: "explainer_storyboarding"
description: "Turn an idea into an explainer video plan: write a voiceover script at ~150 words per minute, pair every line to visuals in a two-column A/V script, sketch storyboards, and lock timing with an animatic before animation starts. Use when asked to storyboard an explainer, script a product overview video, or plan visuals that sync to narration. Hands off to explainer_animation for the render."
---

# Explainer Storyboarding

## Purpose
Turn a raw idea into a locked pre-production package for an explainer video: the goal, the script, the visual plan, and a timed animatic. This covers everything up to animation. The render itself belongs to `explainer_animation`.

## Workflow
1. **Lock the goal in one sentence.** "After this video, the viewer understands X and does Y." If you cannot finish that sentence, you are not ready to write.
2. **Draft the script out loud, then cut to word count.** Planning math: about 150 words per minute of spoken English, comfortable range 140-160. So 60 seconds = 140-160 words, 90 seconds = 210-240 words, 2 minutes = 280-320 words. Read the draft aloud with a timer. Any word you stumble on gets rewritten; any spot that runs long gets trimmed.
3. **Structure the script.** Proven order for product explainers: hook in the first 3-5 seconds, problem, solution, 2-3 key features in action, brief social proof, one clear call to action. Keep it conversational, short sentences, contractions. Write for the ear, not the eye.
4. **Build the two-column A/V script** (template at `assets/two-column-script.md`). Every line of voiceover gets a paired visual. A line with no picture gets rewritten or cut. The visual column must be actionable, "character walks in from left, stops at laptop", never vague, "show the product". One idea per shot: a new idea means a new visual. On-screen text reinforces key numbers only, never full dialogue.
5. **Storyboard.** Rough rule: one frame per sentence of narration, which puts a 60-120s video around 12-20 frames. Each frame carries the key visual, character and action positions, entrances, exits, transitions, and a timing note. Document motion direction or the animator guesses. This is the last cheap point for structural change, so get sign-off here.
6. **Make one style frame.** A single key scene in full color and brand treatment. It defines the look for every asset downstream.
7. **Build the animatic.** String the storyboard stills over a scratch voiceover read plus placeholder music, and watch it at full speed. Check pacing: does each scene land on its words, does anything drag, does the hook grab in the first seconds? Revise the storyboard here, never in animation.
8. **Lock timing with real audio.** Record (or synthesize) the actual voiceover before animation begins, because audio length sets the final video length. Feed the locked scene durations to `explainer_animation` for the render.

## Output Contract
- One two-column A/V script with word count and estimated runtime on every row.
- One storyboard: frames with visuals, motion direction, transitions, and per-scene timing notes.
- One style frame reference or written style spec.
- One animatic pass plus written timing notes: what changed and why.
- Everything downstream-ready: scenes, durations, and VO audio for the animator or the render pipeline.

## Operating Rules
1. Verify before asserting. Every factual claim about the subject must be checked against a real source. No invented numbers, dates, quotes, or events. Unverifiable content gets cut.
2. If quality and cadence conflict, quality wins. A weak script or storyboard never ships to animation. Rework it.
3. Never lock storyboards from memory of the script. Re-read the final script when building frames; the two must agree line for line.
4. The voiceover sets the clock. Do not finalize scene durations from word-count estimates alone; confirm with the actual recorded read.
5. Plain voice everywhere, no em dashes, no AI-marketing phrasing, in scripts and on-screen text.