Story Design

This file defines the story: what the video teaches, in what order, and why each frame exists. It does not define layout, visual effects, animation, or final markdown schemas. For exact file syntax, follow ../hyperframes/references/storyboard-format.md and ../hyperframes/references/script-format.md.

Read first

Read these inputs before writing:

1. hyperframes.json — locked brief: angle, length, aspect ratio, language.
2. frame.md — tone, mood, design system, and register.
3. capture/extracted/visible-text.txt — the article / notes / topic / brief (the source of information).
4. user_script.txt and VO_MODE, when the user pasted a script.

There is no asset-descriptions.md and no capture/assets/ to inspect — this is faceless. Every visual is invented downstream (Steps 4-5); your job here is the narrative, not a visual asset list.

Output

Create two files:

• STORYBOARD.md — the teaching plan, one frame per beat.
• SCRIPT.md — the locked narration, only for spoken frames.

Every storyboard frame must include the required fields from the storyboard format reference, plus the narrative metadata below.

Core rule

An article is an information dump. A video is a guided act of understanding.

Do not follow paragraph order. Reorder, merge, omit, and compress the source text into a clear teaching sequence. Strip the asides; surface the spine. The single most common failure is paraphrasing the article in order — do not do that. The input text is the source of information, not a story template.

Step 3 method

1. Extract the teaching truth

From the brief and text, identify:

• Audience — who the video is speaking to, and what they already (don’t) know.
• Gap or stakes — the confusion, question, or “why care” the explanation resolves.
• Thesis — the one-line idea the viewer should walk away with.
• Spine — the 3-6 ideas (mechanisms / steps / items / beats) that build to the thesis.
• Evidence — the concrete numbers, examples, comparisons, or worked cases that ground it.
• Landing — the takeaway or the call to think / try / act.

Write the storyboard around the thesis, not around the article’s sections.

2. Match the register to frame.md

Use frame.md as a soft guide — the visual system tunes the voice, not the structure:

The teaching truth decides the arc. The visual system tunes the voice.

3. Choose one explainer structure

Pick one structure (or explicitly name a compound). Do not splice phases from different structures — each is a complete path through understanding.

Choosing: one idea to understand → concept; an ordered procedure → how-to; parallel co-equal items → listicle; a concrete narrative/case → story.

Compounds layer an outer arc with an inner rhythm — e.g. concept-explainer with process (ordered steps inside the mechanism phase), story-explainer with how-to. Set arc in the frontmatter to the chosen structure (or <outer> with <inner>). The downstream visual phase reads it for pacing: a process inner rhythm means tighter seams on a consistent stage and shorter frames.

4. Build the frame sequence

Each frame needs one clear job. Avoid frames that only say “more detail” or “another point.”

For every frame, define (use the storyboard format’s fields, with these narrative additions in the frame’s metadata + prose):

• type — one of hook | pain_point | product_intro | feature_showcase | benefit_highlight | social_proof | branding | cta. This enum is shared with the downstream visual layer for pacing; repurpose it for teaching per the mapping below.
• persuasion — a named rhetorical / clarity technique (see catalog), not “explain the idea.”
• beat — the target feeling (see vocabulary).
• scene — a one-line visual idea, not detailed composition.
• voiceover — spoken guide text, or empty for silent frames.
• transition_in — a registry transition name (see Transitions).
• blueprint (optional candidate) — consult the role→blueprint menu in ../hyperframes-animation/blueprints-index.md; when a proven shape fits this beat, tag its id (a tag, not a commitment — Step 4 confirms or overrides). Then write the voiceover in the shape that blueprint implies, so the line is reveal-ready before Step 4 ever runs. Teaching truth still decides which beats exist — never invent, drop, or bend a beat just to fit a shape; omit blueprint and write the line plainly when none fits.

In the prose under each frame, state:

• narrativeRole — the scene’s job in the explanation (e.g. “Concretizes compound interest as a snowball,” not “Shows a chart”).
• keyMessage — the one thing the viewer should understand after this frame (one sentence).

Type-enum repurposing (shared enum → explainer roles)

The enum is shared with the downstream visual layer; map your explainer roles onto it so downstream pacing matches the frame’s job:

The body is usually a run of feature_showcase (steps/mechanisms/items), interleaved with benefit_highlight (implications) and social_proof (examples/data). At least one feature_showcase or product_intro should exist (every explainer has a body and a named idea).

Hook strategy

Pick one opening strategy for the first 3-5 seconds. For explainers the hook opens a cognitive gap or stakes:

The hook must create curiosity, tension, or stakes. Do not open with a generic definition. Per ../hyperframes-creative/references/story-spine.md: the hook speaks the viewer’s language (the payoff of understanding, never the source text’s section headings), and the thesis (message) lands by beat 2 — the explanation after that is its evidence.

Clarity / rhetoric technique catalog

persuasion is a named technique — how this frame makes the idea land or clear — not a vague intent. Combine when several are active (e.g. “Analogy + progressive disclosure”).

When no catalog technique fits, name a new one inline and explain its mechanism (e.g. “Subtractive framing: define the concept by what it is not first”). Never write generic “explain the idea.”

Emotional beats

beat is one word or a short compound phrase (e.g. “Curiosity and clarity”). Avoid generic “positive” / “interested.” Explainers ride a comprehension arc:

• Negative valley — open the gap (hook / pain_point): curiosity · puzzlement · surprise · tension · concern · skepticism · recognition · intrigue
• Pivot — orient (product_intro / concept-naming): clarity · orientation · anticipation · focus
• Build — build understanding (feature_showcase / benefit_highlight / social_proof): comprehension · “aha” · confidence · fascination · foresight · momentum · conviction · delight · unease (for a caveat) · mastery
• Resolution — land (branding / cta / final): clarity · satisfaction · resolve · inspiration · inevitability · “now I get it”

Compound beats are often strongest, e.g. “Surprise + recognition”, “Comprehension + delight.”

The body is a sequence, not a single frame

An explainer’s core is almost always 3-6 body frames on a consistent visual stage, each advancing one mechanism / step / item / layer, building understanding cumulatively. A single isolated body frame rarely teaches anything.

• concept-explainer: name the concept (product_intro) → reveal the mechanism layer by layer (a run of feature_showcase, interleaving benefit_highlight for “so what” and social_proof for a grounding example).
• how-to-process: feature_showcase per step, ordered, on one stage. Carry the object being acted on across adjacent steps (see Continuity).
• listicle: feature_showcase per item; items are parallel, so default to cut / push-slide between them.
• story-explainer: frames follow the beats (setup / tension / turn / resolution / lesson); types map per the table (pain_point for tension, branding for the lesson).

Continuity across frames (no worker grouping)

This framework builds one frame per worker — there is no “continue run” that hands several frames to one worker. A sequence of frames reads as one continuous shot through two storyboard-level levers, both yours:

1. A consistent stage — consecutive body frames share the same composition idea (same diagram growing, same number line, same desk), stated in each frame’s scene so Step 4 and the workers keep the stage stable.
2. A consistent transition — pick one seam type for a sequence (usually push-slide <DIR> for ordered steps, crossfade for a soft layer reveal) and repeat it across the run, so the frames feel like one flow rather than separate slides.

When a single element genuinely transforms between two ideas (a diagram node becomes a chart bar, a formula becomes its result), keep it within one frame as a development beat (entrance → the transform → settle) rather than splitting it across a seam — the worker owns that motion. Note the intent in the frame’s scene / narrative; Step 4 turns it into a time-coded shot sequence (instantiating the candidate blueprint).

Transitions

Use only registry transition names in transition_in:

cut | crossfade | blur-crossfade | push-slide LEFT | push-slide RIGHT | push-slide UP | push-slide DOWN | zoom-through | squeeze

Pick 2-3 transition types for the whole video and repeat them. Frame 1 uses cut as a placeholder (there is no previous frame). Match the seam to the narrative: ordered steps → a consistent push-slide; a soft layer reveal or atmosphere shift → crossfade / blur-crossfade; zooming into a detail or pulling back → zoom-through; a clean topic switch or new list item → cut.

Faceless visuals — no asset inventory

Every visual is invented downstream from each frame’s narrativeRole / keyMessage / scene — typography, abstract graphics, diagrams, data-viz are all first-class. Therefore:

• Do not write an asset_candidates line describing intended diagrams or typography as if they were files. Visual intent belongs in scene + narrativeRole; the visual phase reads those.
• The only real asset is a user-supplied image already placed at public/<basename>. Then add one line asset_candidates: public/<basename> — <≤25 words: what it is>. Never invent paths or reference capture/.

Script rules

If there is no pasted script

Write tight per-frame narration:

• 1-2 sentences per spoken frame; usually 6-20 words.
• Concrete and human; teach, don’t read the article aloud.
• Write each line as discrete cues, not one run-on breath. Step 5 reveals each on-screen piece when the voiceover names it (the anti-PowerPoint mechanism). A line with clear phrase boundaries — “First the snowball — then the hill — then the speed” — hands the shot its reveal cadence for free; a single long clause leaves the frame nothing to pace to.
• Strong (concretization): “Compound interest isn’t addition, it’s a snowball — every turn picks up the snow from the last, then more.”
• Weak (article-paraphrase): “The study, published in 2019, examined three cohorts and found that…” — that is reading, not explaining.

Avoid: “Unlock the power of…”, “Seamless experience”, long noun-phrase lists, a frame that is only a filler bridge (“Or…”).

Silent frames are allowed and common in explainers — a diagram assembling itself, a worked example animating, a beat of held tension before a turn. Set voiceover empty and leave the frame out of SCRIPT.md; then narrativeRole + persuasion must carry what the script doesn’t say.

If VO_MODE = restructure

Treat user_script.txt as source material. Rewrite, reorder, merge, or omit to fit the chosen structure and target length.

If VO_MODE = verbatim

Do not rewrite the user’s words. Segment the script into frame-sized chunks at sentence or paragraph boundaries (you may split a long sentence at a natural clause boundary, but do not change words). Final duration follows the provided script.

Music & silence

The storyboard’s top YAML block carries a music: field — the BGM mood the audio step retrieves against (e.g. music: confident minimal tech underscore). Omitting it falls back to message: → arc: → a neutral default, so BGM plays unless turned off explicitly.

• music: none — BGM off (narration, if any, still runs).
• music: none + no SCRIPT.md — the canonical fully-silent marker: no narration, no BGM, no SFX. audio.mjs generates nothing and the audio step is a clean skip. Use exactly this spelling when the user asks for a silent / music-free video.

Frame template

Use the exact fields required by the core storyboard format. This is the narrative shape each frame should satisfy:

Final checklist

Before asking for user approval, verify:

• One explainer structure is named (compound only when explicitly named); the sequence is narrative-driven, not paragraph-order-driven.
• The opening uses a named hook strategy.
• Each frame has one job; the body builds cumulatively (a run of feature_showcase / benefit_highlight / product_intro), not a single isolated body frame.
• Every frame has type, persuasion (a named technique from the catalog), and beat (specific, not generic).
• Each voiceover is phrase-segmented into cues (each a piece Step 5 can reveal on), not one run-on clause; a candidate blueprint: is tagged wherever a proven shape fits, and omitted where none does.
• The emotional arc has meaningful variation matching the structure.
• Transitions use only registry names and repeat 2-3 types; frame 1 is cut.
• A consistent stage + consistent transition carry any multi-frame sequence; a genuine element transform stays inside one frame.
• asset_candidates is absent (faceless) except a real user-supplied public/<basename>.
• SCRIPT.md contains only locked spoken narration; silent frames are intentional and omitted from it.