Visual Design

The unit is a time-coded shot sequence

A frame’s visual layer is a sequence of time windows paced to the voiceover, not a bag of effect tags. The failure that reads as PowerPoint is front-loading: the agent rushes the whole canvas on screen in the first ~25%, and then it just sits. A time-coded shot sequence written against the VO makes that impossible: each window states what is on screen and what is moving, and nothing appears before the voiceover reaches it. In a PR explainer the development often is the reveal — the diff hunk typing in, the before→after morph, the request-retry diagram running, the impact stat landing. Let the build be the message.

Write each frame as a handful of windows cued by the spoken line:

• Each Scene line names what’s on screen, what moves in this window, and where it sits (layout, inline). Times are real seconds across the frame’s duration.
• Pace reveals to the voiceover; never front-load. This is the core anti-PowerPoint mechanism (→ motion-language.md Part 2 Rule 2). At t=0 show only what the VO is saying then; reveal each further piece — a line, a file chip, the hunk, a stat — when the VO names it, spreading reveals across the shot and especially the back ~50%. The window count = the number of spoken cues the line calls for. There is no fixed count and no mandatory “middle” act; the only sin is dumping everything up front.
• End on a held read. Once the content has resolved it holds and reads — prefer stillness to bad motion: no forced camera drift, no lazy breathing, no back-half pan/push; at most a subtle jitter keeps it alive (→ motion-language.md). Only the final frame has a real exit; every other frame’s exit is the harness transition (story’s transition_in).
• A deliberately held frame — content already revealed, now reading still — is legitimate and often right (a climax, a breather). The failure is never “too still”; it is front-loaded-then-frozen. Place held beats deliberately for rhythm (allocate them in ## Video direction).

Pick the shape — instantiate a blueprint

Don’t invent each shot from scratch. The frame’s role (its type / beat) points to a proven shape:

1. Match the role to a blueprint. Open ../hyperframes-animation/blueprints-index.md, find the frame’s role in the role→blueprint menu, and pick the blueprint whose intent fits this beat (story may already have named a candidate id — confirm or override it). Read that blueprints/<id>.md: it is a short, domain-agnostic, time-coded shot template with [slots] and a named signature move.

2. Instantiate its [slots] with THIS frame’s content — three postures:

   • Reproduce — the blueprint fits the beat and your content maps onto its slots cleanly. Fill every [slot] and follow its Scene timing.
   • Adapt — the structure fits but the content / surface doesn’t. State what you keep / what you change in one line, then write the adapted Scene lines. You may never drop the signature move, and you keep the reveals paced to the VO.
   • Compose — no blueprint fits the beat. Build the shot from the motion vocabulary in motion-language.md: still pace reveals to the VO. Mark it blueprint: compose.

3. Keep the signature move. Whichever posture, the blueprint’s signature move is the spine of the shot — carry it through.

A code beat is the one place you don’t pick a blueprint for the centerpiece — the code-* block is the shape (see PR code beats below). You still write the Scene sequence for the surrounding surface.

What you add to each frame

Story-design’s ## Frame N block already carries the narrative. You append the shot. Story’s scene / voiceover / transition_in / role fields stay untouched.

The lightweight tags:

• blueprint: — the id you instantiated (with (Reproduce) / (Adapt)), or compose. One id per frame.
• focal: — for a concept/mechanism beat, the invented hero (a hero word, a diagram, a number-lockup); for a code beat, the code-* block (name the block + the hunk); for the credits close, the avatar row.
• roles: — each element’s role: foreground subject · background (full-bleed, dim 30–50%) · supporting. Invented elements you design; the only real assets are the credits assets/<login>.png avatars (named in story’s asset_candidates).
• sfx: — name the sound the beat wants; the audio script’s fetch-sfx retrieves it and the assembler mounts it at root — you only name it, never embed <audio>.

Layout + motion are stated INLINE in each Scene line — name the template / density / depth as part of “where it sits”, and name the move from motion-language.md‘s vocabulary; let it settle on a long-tail curve (power3 default). Never write px / scale / ease curves / ms (the worker writes those).

PR code beats — name a code-* block

For a diff / before_after / code beat, the frame's centerpiece is a ready-made code-* registry block, not an invented HTML visual — the one exception to “invent every visual.”

• Name the block in scene + focal. Pick the one that fits the beat (before→after = code-diff; refactor/rename = code-morph; new code written on = code-typing; spotlight a line = code-highlight; walk a long file = code-scroll; a hero reveal = code-3d-extrude / code-particle-assemble). Full map → code-vocabulary.md. Name the hunk too (“the request() retry block, ~6 lines”). The block is the focal; the Step-5 worker installs + fills it with the real diff.
• The block owns the code animation; your Scenes choreograph the surrounding Code Surface. The block is the development beat (the diff/typewriter/morph plays on its own cadence — the worker only fits it to the frame’s data-duration so a long snippet doesn’t overrun). Your Scene windows move the code-editorial Code Surface around it: the navy window seating in, the file header typing on, the camera settling onto the hunk, a +N/−M count-up, a coral underline drawing on the landed line. Name those moves inline; do not re-specify the code animation itself. A code beat is usually blueprint: compose (the block is the shape).

PR mechanism beats — invent an animated diagram of the behavior

A mechanism frame is the show-the-behavior beat — the antidote to a video that only shows code + text. Its focal is an invented animated diagram that plays out what the change does at runtime (the request retrying, the cache filling, serial→parallel, the race resolved) — not a code-* block and not a headline.

• Name the behavior + the diagram in scene + focal. e.g. scene: "animate the request lifecycle — fire → 500 → backoff → retry → 200, invented SVG flow"; focal: the request-lifecycle flow. Reach for the flowchart / flowchart-vertical / data-chart registry blocks where they fit (name them in scene so Step 5 pre-installs them); otherwise the worker builds it in SVG / HTML / GSAP from code-editorial’s atoms.
• The build IS the shot sequence. Unlike a code block (which owns its own animation), the diagram is yours to choreograph across the Scene windows — the lanes / nodes draw on (Scene 1), the flow runs / the lane splits / the front advances as the VO names each step (middle Scenes), the resolved state + one coral emphasis lands (final Scene). Never let it enter then freeze.
• Stay on code-editorial’s cream ground, hairline-ink. Nodes / edges / lanes in hairline ink on cream; one coral marker on the active or changed element; mono labels. Not the navy code surface (that’s for code), not heavy shapes / bokeh. Plan it into the top ~83% (caption keep-out).

A mechanism frame carries no asset_candidates (it’s invented, like every non-credits frame).

Impact & credits

• Impact / evidence — numbers (+1,204 / −318, files touched, perf delta) go on an impact frame as a number-lockup (code-editorial’s Number/Impact treatment): name it the focal, reveal it with a count-up paced to the VO.
• Credits close — the optional credits frame uses the real assets/<login>.png avatars (named in story’s asset_candidates) as the focal: an avatar row that staggers in. This is the one frame with non-empty asset_candidates and real assets.

Inventing the visual (non-code beats)

Every non-code, non-credits beat (hook / change / cta / concept) is designed, not captured. Three first-class treatments:

• Typographic / kinetic type — a hero word, the PR’s headline claim, a stat. Treat type as the subject: full-bleed scale, weight contrast, one emphasized term. Strongest for hooks and the cta.
• Abstract graphics — shapes / paths / geometry that embody the idea the script names; don’t decorate with generic bokeh.
• Diagram / data-viz — the mechanism diagrams above, a data-chart for a perf delta, a number-lockup. The build (each part on beat) is the teaching — design it to assemble across the Scenes.

Make the invented hero fill 40–60% of the frame — big enough to read; don’t shrink the one designed element into decoration around empty space.

Layout — named inline per Scene

State each Scene’s layout as part of “where it sits.” If the blueprint (or the code-* block) already implies a composition, that wins — describe it directly; the vocabulary below is for composing freely. Never write px / scale / shadow (the worker does). One frame’s layout can EVOLVE across its Scenes. Use ≥3 different framings per video; never the same framing twice in a row.

• Framing vocabulary — centered (hero / climax / a single code surface) · rule-of-thirds · split-screen (before/after, two surfaces) · layered-depth (immersive) · asymmetric 60/40 or 70/30 (a code surface + a caption rail) · triptych (three changes at once) · full-width strip (a file list / timeline). Let the beat decide, not a quota.
• Density — primary visual ≥ 40% of canvas; ≥ 3 depth layers; never a lone small cluster floating in empty space. Openings/closings are prone to emptiness — add environmental layers (a dim grid, low-opacity scanlines, brand-color ambient). Squint test: after blur you can still pick out the #1 element.
• Hierarchy — combine ≥ 2 of size (3:1) / weight (800 vs 400) / contrast / position (upper-third is golden) / motion, so one element clearly dominates.
• Depth — layer 2–3 of: size, blur, opacity gradient, overlap, shadow-stack, counter-scale on a push.
• Don’t show: nav bars, footers, scrollbars, real cursors / browser chrome, generic decorative shapes, floating bokeh / purple-blue “AI” gradients (banned). The navy code surface is for code beats only; mechanism diagrams stay on cream.

Portrait & square (non-16:9 canvases)

The zones, density, hierarchy, and depth principles all still apply; the aspect ratio changes, and a wide layout doesn’t transplant into a tall one — design for the storyboard’s format from the start.

• Stack vertically, not side-by-side — split-screen / triptych / 60-40 become top/bottom stacks. A code surface runs nearly full-width in portrait with fewer visible lines.
• Vertical center moves with the canvas — anchor a centered hero around y ≈ 0.42 × height (portrait ≈806, square ≈454), not a fixed 540.
• Type runs larger, fewer words per line. Travels well to portrait: Centered, Layered Depth, Full-Width Strip; avoid wide Split Screen / Triptych — use stacked equivalents.

## Video direction — write the invariants ONCE

The whole video shares one look and one motion grammar. Write a ## Video direction block ONCE at the top of STORYBOARD.md so every frame inherits it and per-frame Scene lines carry only the delta. This block is load-bearing — keep it.

• palette system — from frame.md (code-editorial): which roles map to which hues. Never invent.
• motion grammar + reveal model — long-tail eases (power3 default, smooth over bouncy) + the VO-paced reveal model + what may stay alive during a hold (subtle jitter at most) (→ motion-language.md).
• rhythm / held-frame allocation — name the held / breather frames so the video varies its energy.
• negative list — off-brand textures, plus both motion failure modes — slideshow (front-load then freeze) and screensaver (everything floating independently) (→ motion-language.md).

Do not repeat these per frame.

Palette & type — from frame.md, never invented

• Palette — frame.md (code-editorial) is the color truth; apply its roles per frame. Generic basics → hyperframes-creative/references/house-style.md.
• Type — fonts resolve via frame.md‘s type tokens; reference them by role (display / body / mono / the pack’s ramp), never by raw family or px. Code surfaces and mechanism labels use the mono role. Typography craft → hyperframes-creative/references/typography.md.

Caption-band keep-out (plan side)

The bottom ~17% of the canvas is reserved for the caption pill. Plan every frame’s content into the top ~83% (the worker enforces the pixel cutoff). When captions are enabled, primary content caps at the band top, and a centered hero anchors at y ≈ 0.42 × height (landscape ≈454, portrait ≈806); background / ambient layers are exempt and may stay full-bleed. Holds even when captions are disabled — bottom-edge consistency.

Where the detail lives

Before you finish — checklist

• ## Video direction written once at the top (palette · motion grammar + shot model + idle budget · stillness allocation · negative list incl. both failure modes); per-frame entries are deltas.
• Every frame is a time-coded shot sequence with real second windows across its duration — not a tag bag.
• No frame front-loads — at t=0 only what the VO is saying enters; each further piece reveals on its spoken cue, across the back ~50%. Window count follows the VO.
• Every frame names a blueprint: id (Reproduce / Adapt) or compose; an Adapt keeps the signature move; nothing collapses to a single front-loaded dump.
• Code beats name a code-* block as the focal, let the block own the code animation, and choreograph only the surrounding Code Surface in the Scenes.
• Mechanism beats name an invented animated diagram of the behavior (or a flowchart / data-chart), choreographed across the Scenes on code-editorial’s cream ground with one coral marker — not a code block, not typography; the body is not an unbroken run of code surfaces.
• Impact uses a number-lockup with a count-up; the credits close uses the real avatars as the focal.
• Each non-code, non-credits frame names its invented focal + per-element roles, kept few and load-bearing.
• Layout + motion named inline per Scene (no px / ease curves / ms / JS).
• Content planned into the top ~83% (caption band clear); palette / type pulled from frame.md by role.
• You wrote no HTML.