Businesses in ChatGPT Launch Tour
A 49-second, music-led launch film that alternates bold benefit cards with concise UI proofs of setup, customer help, and insights.
By OpenAI for OpenAI / ChatGPTSource onlineOriginal post ↗
Copies get_blueprint("REF-6241") for Claude. Connect Claude
Why it works
The opening promise is legible within the first 3 seconds while commerce props and changing names add motion around stable central type. Five hidden hard cuts at frames 200, 245, 682, 972 and 1240 divide a single indexed shot into a clear claim-to-proof sequence; 34 of 49 measured cues fall on beats, carrying a nearly wordless film.
Blueprint
15 plates that take the film apart, from its format to a recipe you can run. Claude reads the same plates over MCP.
Plate 01
The format in one breath
A clean, music-led announcement turns a business claim into a guided tour of useful actions. It alternates large declarative type with actual interface proof, then closes on one imperative.
Why it works
- The first 3 seconds make the premise legible before any interface detail.
- The five measured hard cuts at frames 200, 245, 682, 972 and 1240 separate jobs without breaking the light, consistent canvas.
- The beat grid supplies momentum while small UI changes carry information; 34 of 49 recorded cues are on beats.
Plate 02
Format card
| Platform · aspect · length | YouTube · 16:9 · 48.7 s · 640×360 source · 30 fps |
| Pace | One indexed shot, five hard cuts inside it; long UI holds with animated details |
| Script | Music-led · effectively wordless; on-screen copy delivers the argument |
| Voice | None |
| On screen | White canvas, animated type, business/profile cards, product feed, mobile agent, analytics, end card |
| Type | Three cap-height levels: 68–77, 34–47, 27–29 px; sentence case, short builds |
| Sound | Tonal electronic pulse · 136 BPM measured · cue accents · reference −22.1 LUFS |
| Structure | Six beats · product proof in the middle · call to action after frame 1240 |
Tools you'll need: Remotion for motion graphics · screen recordings of the product · a licensed or generated music track · Claude alone: partly
Frame sheets
Frame by frame
Stills taken at measured moments, each labelled with its time. Claude gets the same sheets over MCP.
Plate 03
The hook, frame by frame (first 3 seconds)
| Time | Picture | On-screen text | Words and sound | What it does |
|---|---|---|---|---|
| 0–0.5 s | Blank light field; small commerce objects enter from edges | First line begins | Whoosh at frame 6 | Creates room for a simple claim |
| 0.5–1.5 s | Objects frame the centre without covering it | Premise completes | Hit at frame 20 | Makes the message readable in a glance |
| 1.5–3 s | Objects leave; brand names cycle between two fixed lines | Changing business examples | Tonal accents | Shows that the offer applies to many businesses |
Hook formula (inferred): Your {{PRODUCT}}, now with {{FEATURE}}.
Why it holds (inferred): The centred sentence stays spatially stable while the examples change. The opening is a typographic promise before it asks viewers to inspect UI.
Plate 04
Structure (beat sheet)
| # | Time (s) | Shots | Beat | What happens | Technique |
|---|---|---|---|---|---|
| 1 | 0–6.67 | 1 | Hook | Premise and rotating examples resolve into a business card | Edge props; central type; card build |
| 2 | 6.67–14.3 | 1 | Claim and maintain | Claimed state leads to editable business information | Hard cut at 200; text bridge at 245; vertical form travel |
| 3 | 14.3–22.73 | 1 | Add inventory | Upload control, feed preview and populated storefront | UI builds and card replacement |
| 4 | 22.73–32.4 | 1 | Answer and browse | Agent configuration becomes a customer conversation and product view | Hard cut at 682; mobile panel enlargement |
| 5 | 32.4–41.33 | 1 | Understand activity | Title resolves to metric cards, chart and editable profile | Hard cut at 972; staggered data build |
| 6 | 41.33–end | 1 | Call to action | Central imperative, edge objects, verification mark, final wordmark | Hard cut at 1240; restrained exit |
Retention map (inferred): re-hooks at the large type bridges around 8.17, 20.9 and 32.4 s; payoff is the mobile answer and analytics in beats 4–5; call to action is beat 6; ending is a hard stop with no fade to black.
For agents: Shots
The measured data and build specs agents work from. Over MCP they get it all, with the full timeline as JSON.
Shots (measured)
One v1 shot with 5 hidden hard cuts inside it (see "changes inside"), 48.72 s = 1461 frames at 30 fps. Frames are at 30 fps, first and last inclusive: a shot "33–70 · 38 f" is a Remotion <Sequence from={33} durationInFrames={38}>; at another frame rate, use the seconds.
| # | in–out (s) | frames @30 (first–last · count) | dur (s) | transition in | changes inside | scales (inferred) |
|---|---|---|---|---|---|---|
| 1 | 0.00–48.60 | 0–1457 · 1458 f | 48.60 | start | f200 cut · f245 cut · f682 cut · f972 cut · f1240 cut | scales |
scales: fixed = keep this shot's length when the film gets longer or shorter · scales = stretch or shrink it with the length · optional = drop it for the 15 s cut.
Measurement notes: 5 hidden cuts / dips / reframes found inside v1 shots (luma-normalised frame difference): S1 cut at 6.667 s (f200), S1 cut at 8.167 s (f245), S1 cut at 22.733 s (f682), S1 cut at 32.400 s (f972), S1 cut… · S1: 159 text items read; the 40 largest are kept · cue sheet keeps 49 of 176 sound events (all near cuts, then the strongest)
Frame sheets (measured): shots, one labelled frame per shot (labels S<n> m:ss.s; + = just after a change inside the shot) · the full-length timeline, the whole film at fixed steps (labels S<n> m:ss.s f<frame>), 5 pages · close-up strips, 14 frames around each cut and change: f200 cut, f245 cut, f682 cut, f972 cut, f1240 cut (sheet "transitions", 2 pages; labels S<n> m:ss.sss f<frame>): call get_frames('yt-Z9k5oI3Paoc'), or see them on the film page (Pro, or free films).
Plate 05
Shot grammar
Mix: Mostly 2D interface demonstrations and full-frame type, with small dimensional commerce props in beats 1 and 6. Rhythm: The indexed shot is continuous only in metadata. Five measured internal hard cuts establish six visual phases; builds and replacements occur between them.
Rules this editor follows (inferred):
- Let the important title finish before introducing dense UI.
- Keep the camera plane fixed; move cards, words, cursor and props inside it.
- Use the same light field across hard cuts so the subject, rather than a background change, signals the new beat.
For agents: Camera and motion · Transitions
The measured data and build specs agents work from. Over MCP they get it all, with the full timeline as JSON.
Camera and motion (measured)
Per shot, measured on the frames. "Camera move" names the camera's move; the numbers are the image's motion: pan in % of frame width (x) or height (y) per second, +x = content moves right (so a pan left), +y = content moves down; zoom per second > 1 = content grows (push in). "Over the shot" is the total, with px at 1920×1080. Speed is the move through the shot, about one character per 0.5 s (▁ still → █ its peak), then the peak in % of width per second. Cadence: "ones" = a new picture every frame, "twos" = each picture held two frames. Blur is relative to this film (0 = its sharpest shots, 1 = its blurriest); motion = share of the frame changing, 0–1. "?" = confidence under 0.5: a hint, check the sheets. The labels describe how the image moves, not where a 3D camera went: an object flying toward a still camera can read as a push in.
In 1 of 1 shots the camera holds still and something inside the frame moves: see Element motion right under this table.
| # | camera move | pan x · y (%/s) | zoom (/s) | over the shot | ease | speed · peak | cadence | blur | motion | conf |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | static? · element moves ↓ | +0.0 · +0.0 | ×1.000 | none | hold | ▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ | ones | 0.61 | 0.05 | 0.37 |
Element motion (measured): in these shots the frame holds still and something inside it moves: a card, a UI panel, a product, a word. Region = where it moves (x · y · w · h % of the frame); numbers as above, for the moving region; "moving" = share of the shot in motion; ↳ = the move changes direction.
| # | region x·y·w·h % | element move | x · y (%/s) | zoom (/s) | total while moving | ease | speed · peak | moving | conf |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 36 · 49 · 27 · 5 → 46 · 19 · 8 · 13 | slide up | −4.5 · −19.1 | ×0.941 | x −106.8% (−2051 px) · y −449.6% (−4856 px) · ×0.24 scale · roll −89.1° in 23.6 s | irregular | ▂▁▁▁▂▁▁▁▁▁▁▁▁▂▃▁ · 141%/s | 0.49 | 0.75 |
| ↳ 0.07–0.40 s: slide up | +0.0 · −25.5 | ×1.000 | y −8.5% (−92 px) | ease-in | |||||
| ↳ 0.40–1.73 s: scale down | −1.3 · −2.6 | ×0.609 | x −1.7% (−33 px) · y −3.4% (−37 px) · ×0.52 scale · roll +5.5° | ease-out | |||||
| ↳ 1.73–2.07 s: slide down | +0.7 · +8.7 | ×1.007 | y +2.9% (+31 px) · roll +0.6° | ease-out | |||||
| ↳ 2.07–3.33 s: slide up | +0.0 · −20.6 | ×0.996 | y −26.1% (−281 px) | irregular | |||||
| ↳ 3.33–3.60 s: slide right | +12.8 · +0.1 | ×0.998 | x +3.4% (+66 px) | ease-in | |||||
| ↳ 3.73–4.07 s: slide up | +0.1 · −16.5 | ×0.995 | y −5.5% (−59 px) | ease-out |
Focus (measured): shot 1 0.00–6.67 s: sharp inside 0 · 0 · 100 · 100, no defocus (deep focus) · shot 1 6.67–8.17 s: sharp inside 17 · 14 · 67 · 86, no defocus (deep focus) · shot 1 8.17–22.73 s: sharp inside 17 · 0 · 67 · 100, no defocus (deep focus) · shot 1 22.73–32.40 s: sharp inside 8 · 0 · 83 · 100, no defocus (deep focus) · shot 1 32.40–41.33 s: sharp inside 8 · 0 · 92 · 100, no defocus (deep focus) · shot 1 41.33–48.60 s: sharp inside 8 · 0 · 83 · 71, no defocus (deep focus). Box = x · y · w · h % of the frame. In 2D, blur the layers outside the box by that many px; in 3D, set the focus distance on the box and open the aperture until the background blur matches. Locked-off shots: 0 of 1; 1 more could not be tracked (shown as "static?" or "not measured": check the sheets). Animation cadence: a new picture every frame (on ones): 28.2 unique frames per second of 30 while things move (held-frame share 0.06, 461 moving frames). Match it, or the motion will feel smoother or choppier than the reference.
Transitions (measured)
One v1 shot (shot numbers stay those of v1), with hidden cuts and changes inside it; the measured ones:
| shot | at (s) | frame @30 | change |
|---|---|---|---|
| 1 | 6.67 | 200 | cut: hard cut that v1 did not split on |
| 1 | 8.17 | 245 | cut: hard cut that v1 did not split on |
| 1 | 22.73 | 682 | cut: hard cut that v1 did not split on |
| 1 | 32.40 | 972 | cut: hard cut that v1 did not split on |
| 1 | 41.33 | 1240 | cut: hard cut that v1 did not split on |
What shot 1 shows (inferred from the sheets)
| # | Type | Framing | What's on screen | Text / overlay |
|---|---|---|---|---|
| 1 | Composite sequence | Central title to central cards, then wider dashboard and end card | Hooks, claim UI, editable information, catalog, agent, analytics | Large chapter lines alternate with UI labels |
(inferred) The full timeline shows a title-to-metrics picture change around 34.2 s and a metric-card-to-chart build around 36.5 s; the transition table lists only the hard cuts, so treat these as changes inside shot 1.
Plate 06
Visual style guide
- Look (inferred): High-key, nearly white field and fine grey card boundaries; dark text and a single brighter action colour. The reference uses a mostly white palette; replace it by role with
{{BRAND_BG}}and{{BRAND_ACCENTS}}. - Graphics and overlays (inferred): Small retail objects at the edges, fine bordered cards, gentle shadows and cursor actions. Keep product images supplied in
{{UI_ASSETS}}. - Transitions (inferred): Five instantaneous full-frame hard cuts (plate 5); all other picture changes are element animation or content swaps.
- Branding and end card (inferred): Centred imperative followed by
{{LOGO}}and{{CTA_URL}}in text. The reference ends on its wordmark rather than a fade to black.
At a glance
- Type scale (cap px): 68–77 · 34–47 · 27–29 (3 levels) among the text above.
- Phone minimum: cap height ≥ 2% of frame width = 38 px at 1920×1080
- Whole film (time-weighted):
#FEFEFE93% ·#D8D7D64% ·#AF9C8F1% ·#6B69681% ·#4082EA1%. - Curve facts: average 0.97 · darkest 0.92 at 7.8 s · brightest 1.00 at 0.0 s.
For agents: Type · Type spec · Palette by shot · Brightness curve · Layout boxes · Material card
The measured data and build specs agents work from. Over MCP they get it all, with the full timeline as JSON.
Type (measured)
Sizes are cap height (the height of a capital letter) in px at the 1080p deliverable, 1920×1080. Font size ≈ cap ÷ 0.71 for a typical sans (Inter 0.73, Geist 0.71, SF Pro 0.70, Instrument Sans 0.72): check your own font's cap-height ratio. Cap heights are estimated from text-recognition boxes (median error about 11% on real text; text under about 13 px at 1080p and non-Latin scripts are not read). Box = x · y · w · h as % of the frame, from the top-left.
| text | shot | on screen (s) | cap px | ≈ font px | cap % of height | box x·y·w·h % | position |
|---|---|---|---|---|---|---|---|
| "now in ChatGPT" | 1 | 0.5–3.5 | 77 | 108 | 7.1 | 31 · 55 · 38 · 8 | centre, lower third |
| "Claim this business" | 1 | 4.5–6.5 | 29 | 41 | 2.7 | 33 · 83 · 16 · 3 | centre, bottom |
| "You've claimed Colorful Clatter" | 1 | 7.0–8.0 | 40 | 56 | 3.7 | 32 · 40 · 35 · 4 | centre, middle |
| "Keep your" | 1 | 8.5–10.0 | 70 | 99 | 6.5 | 38 · 40 · 24 · 10 | centre, middle |
| "info up to date" | 1 | 9.0–10.0 | 72 | 101 | 6.7 | 33 · 51 · 34 · 10 | centre, middle |
| "Contact" | 1 | 11.0–12.5 | 34 | 48 | 3.1 | 30 · 51 · 10 · 4 | left, middle |
| "Hours" | 1 | 12.0–14.0 | 36 | 51 | 3.3 | 32 · 40 · 7 · 4 | left, upper |
| "PT Upload a product data file" | 1 | 14.5–16.0 | 47 | 66 | 4.4 | 23 · 51 · 37 · 7 | centre, middle |
| "Add products to ChatGPT" | 1 | 14.5–16.0 | 43 | 61 | 4.0 | 21 · 37 · 37 · 6 | centre, upper |
| "using info you provide" | 1 | 21.0–22.5 | 72 | 101 | 6.7 | 24 · 51 · 51 · 10 | centre, middle |
| "Answer customer questions" | 1 | 21.0–22.5 | 68 | 96 | 6.3 | 18 · 41 · 64 · 9 | centre, middle |
| "Understand how people" | 1 | 32.5–34.0 | 68 | 96 | 6.3 | 22 · 41 · 55 · 9 | centre, middle |
| "discover your business" | 1 | 33.0–34.0 | 75 | 106 | 6.9 | 24 · 51 · 52 · 10 | centre, middle |
| "Top discovery topics" | 1 | 35.0–41.0 | 27 | 38 | 2.5 | 13 · 22 · 20 · 4 | left, upper |
| "15,356" | 1 | 35.0–37.5 | 36 | 51 | 3.3 | 32 · 11 · 7 · 5 | left, top |
| "1,234" | 1 | 35.0–37.5 | 36 | 51 | 3.3 | 45 · 11 · 6 · 5 | centre, top |
| "5,432" | 1 | 35.0–37.5 | 36 | 51 | 3.3 | 58 · 11 · 6 · 5 | centre, top |
| "Claim your business" | 1 | 41.5–45.5 | 72 | 101 | 6.7 | 26 · 43 · 48 · 10 | centre, middle |
Plus 8 smaller or briefer text items, cap 25–52 px, mostly labels inside interfaces: all in the timeline JSON.
Type scale (cap px): 68–77 · 34–47 · 27–29 (3 levels) among the text above.
Left out: 4 reads that are not clean words or were seen in a single frame ("THANK YOU", "DOUG'S BONUTE", "support@colorfulc…"…): blurred, moving or partly hidden text the recognizer misread. They stay in the timeline JSON, not in the type scale.
Phone minimum: cap height ≥ 2% of frame width = 38 px at 1920×1080. Why: in a phone feed the film fills the screen width, about 390 pt, and 2% of that is 7.8 pt, the cap height of 11 pt text, Apple's smallest recommended text size. Words the viewer must read go at or above it; smaller text can only be decoration. Below it in the reference: "Top discovery topics" (27 px, shot 1), "Claim this business" (29 px, shot 1), "15,356" (36 px, shot 1), "1,234" (36 px, shot 1), "5,432" (36 px, shot 1), "Hours" (36 px, shot 1) and 1 more above and 4 of the smaller items. Size these up in your version, or treat them as decoration.
Type spec (inferred)
Assume cap height/font size = 0.71. Use {{BRAND_FONT}}; Instrument Sans is a free fallback only. All essential labels at the small level rise to at least 38 px cap.
| Level | Cap px at 1080p | Font size | Weight · case · tracking | Colour pair | Motion |
|---|---|---|---|---|---|
| Chapter / hook | 68–77 | 96–108 px | 650 · sentence · −0.035 em | dark text on {{BRAND_BG}} | Word reveal over 9–18 f; cubic-bezier(.22,1,.36,1) |
| UI heading / action | 34–47 | 48–66 px | 550–650 · sentence · −0.02 em | dark text on surface | Opacity and y 12 px over 10 f; same ease; raise essential 34–36 px caps to 38 |
| UI detail | 27–29 in reference; 38 minimum in new film | 54 px minimum | 500 · sentence · 0 em | dark text on surface; {{BRAND_ACCENTS}} for action | Opacity over 8 f; decorative microtext may remain smaller |
Palette by shot (measured)
Colours measured on the frames, with their share of the frame; luma 0 = black, 1 = white. These are the reference's brand colours: map them onto {{BRAND_BG}} and {{BRAND_ACCENTS}} by role (background, surface, accent), don't reuse them as the format. Neighbouring shots with the same palette share a row.
| # | luma mean · p95 | colours (share of frame) | highlights (bright pixels) |
|---|---|---|---|
| 1 | 0.97 · 1.00 | #FEFEFE 93% · #D8D7D6 4% · #AF9C8F 1% · #6B6968 1% · #4082EA 1% | #FEFEFE ≈6500 K 66% · #FFFFFF ≈6500 K 34% |
Whole film (time-weighted): #FEFEFE 93% · #D8D7D6 4% · #AF9C8F 1% · #6B6968 1% · #4082EA 1%.
Brightness curve (measured)
Mean luma per 1 s, one character each, grouped in tens; p95 = the bright highlights. Scale: ▁ = 0 (black) → █ = 1.0 (white); luma runs 0–1.
0s 10s 20s 30s 40s
mean ██████████ ██████████ ██████████ ██████████ █████████
p95 ██████████ ██████████ ██████████ ██████████ █████████
Curve facts: average 0.97 · darkest 0.92 at 7.8 s · brightest 1.00 at 0.0 s.
Layout boxes (inferred)
| Element | Shots | Box x · y · w · h (%) | Source · confidence | Notes |
|---|---|---|---|---|
| Hook second line | 1 | 31 · 55 · 38 · 8 | measured text box · high | Keep centred |
| Claimed-state heading | 1 | 32 · 40 · 35 · 4 | measured text box · high | Card text |
| Information heading | 1 | 38 · 40 · 24 · 10 | measured text box · high | Chapter title |
| Catalog upload heading | 1 | 21 · 37 · 37 · 6 | measured text box · high | UI bridge |
| Agent chapter title | 1 | 18 · 41 · 64 · 9 | measured text box · high | Two-line title |
| Analytics title | 1 | 22 · 41 · 55 · 9 | measured text box · high | First line |
| Analytics card heading | 1 | 13 · 22 · 20 · 4 | measured text box · high | Enlarge if essential |
| End-card imperative | 1 | 26 · 43 · 48 · 10 | measured text box · high | Centre line |
| End-card wordmark | 1 | 38 · 44 · 24 · 11 | sheet tile at 48.0 s · medium | Replace with {{LOGO}} |
End card (inferred): {{BRAND_BG}} field; imperative in the measured box above; then {{LOGO}} centred and {{CTA_URL}} as readable text beneath it, both clearing the outer 10% of frame. No fade to black.
Material card (inferred)
:root { --brand-bg: {{BRAND_BG}}; --brand-accent: {{BRAND_ACCENTS}}; --ink: color-mix(in srgb, var(--brand-bg) 4%, black); }
.frame { background:var(--brand-bg); color:var(--ink); font-family:{{BRAND_FONT}}; }
.ui-card { background:color-mix(in srgb,var(--brand-bg) 97%,white); border:1px solid color-mix(in srgb,var(--ink) 9%,transparent); border-radius:22px; box-shadow:0 10px 30px rgba(0,0,0,.06); }
.action { color:var(--brand-accent); }
.prop { filter:drop-shadow(0 12px 14px rgba(0,0,0,.10)); }
No 3D is needed: layer supplied prop PNGs above the type plane, and use CSS perspective only for small edge rotations. Deep focus is measured throughout; 2D blur outside the plate 5 focus boxes is 0 px. Use no bloom or vignette.
Plate 07
Voice and delivery
Measured: The transcript recogniser returned music as a few words. Plate 8 finds speech over only 1% of the film. Voice direction (inferred): No voice. The music and picture carry it; preserve quiet space around title builds.
Plate 08
Sound design
- Music (inferred): Tonal electronic bed with a steady, soft kick and bright bell/pluck accents; lower the density under dense UI, then thin it for beat 6.
- Sound effects (inferred): Align the opening whoosh to frame 6 and title hit to frame 20; use tiny cursor clicks for actual UI actions, under the music.
- Silence and space (inferred): No abrupt silence during the demo; leave an airy tail under the end card. The measured outro is quieter than the steady body.
At a glance
- Speech: words cover 1% of the film (2 words, v1 transcript): little speech.
- Tempo: 136 BPM, confidence 0.99 · first downbeat 0.151 s (bar start confidence 0.27, uncertain) · beat 0.4412 s · bar of 4 = 1.765 s.
- Key: G minor (confidence 1.00; runner-up Eb major).
- Loudness of the reference (measured): −22.1 LUFS integrated, true peak −11.2 dBTP, range 3.1 LU. That is how this film was mastered, not a target.
For agents: Sound grid · Original music bed
The measured data and build specs agents work from. Over MCP they get it all, with the full timeline as JSON.
Sound grid (measured)
Speech: words cover 1% of the film (2 words, v1 transcript): little speech.
Tempo: 136 BPM, confidence 0.99 · first downbeat 0.151 s (bar start confidence 0.27, uncertain) · beat 0.4412 s · bar of 4 = 1.765 s.
Beat grid in frames. Tracked beats, not a constant grid (the tempo drifts): use the listed times, not a formula.
| per beat | per bar | first downbeat (frame) | downbeats (frame), first 8 | |
|---|---|---|---|---|
| 30 fps (this film) | 13.24 f | 52.94 f | 5 | 5 · 60 · 113 · 167 · 220 · 274 · 327 · 380 … |
| 60 fps | 26.47 f | 105.89 f | 9 | 9 · 119 · 226 · 334 · 441 · 548 · 654 · 761 … |
Beats in seconds: 0.151 · 0.656 · 1.097 · 1.544 · 1.991 · 2.432 · 2.879 · 3.320 · 3.767 · 4.214 · 4.673 · 5.120 … (108 in all, audio.tempo.beats in the timeline JSON).
Key: G minor (confidence 1.00; runner-up Eb major).
Sections (by energy of the whole mix, voice included):
| section | time (s) | shots | level (RMS) |
|---|---|---|---|
| steady | 0.0–6.0 | 1 | −22 dB |
| steady | 6.0–11.0 | 1 | −21 dB |
| steady | 11.0–40.5 | 1 | −21 dB |
| outro | 40.5–48.7 | 1 | −24 dB |
Energy map (RMS per 1 s, ▁ = −50 dB → █ = −18 dB):
▇▇██▇▇██▇▇ ███▇▇█████ ▇█████████ ██████▇██▇ ▇▇█▇▇█▆▅▁
Cue sheet (49 cues: 0 at a cut, 34 on a beat; the 24 strongest shown). "On screen" names the picture event the cue hits within ±3 frames (a cut, a jump, a move starting, a line of text appearing, a brightness step), or says there is none:
| t (s) | frame @30 | shot | kind | strength | on screen (±3 f) | on beat |
|---|---|---|---|---|---|---|
| 0.19 | 6 | 1 | whoosh | 1.00 | element slide up starts (shot 1) | ✓ |
| 0.65 | 20 | 1 | hit | 0.99 | "Your business," appears | ✓ |
| 1.09 | 33 | 1 | tonal | 1.00 | no visible event | ✓ |
| 2.88 | 86 | 1 | tonal | 1.00 | no visible event | ✓ |
| 3.32 | 100 | 1 | tonal | 1.00 | element slide right starts (shot 1) | ✓ |
| 13.12 | 394 | 1 | hit | 1.00 | no visible event | ✓ |
| 15.12 | 453 | 1 | tonal | 1.00 | no visible event | |
| 18.88 | 567 | 1 | unknown | 1.00 | no visible event | ✓ |
| 19.34 | 580 | 1 | tonal | 1.00 | element scale up starts (shot 1) | ✓ |
| 20.23 | 607 | 1 | hit | 1.00 | no visible event | ✓ |
| 20.46 | 614 | 1 | tonal | 1.00 | no visible event | |
| 42.45 | 1273 | 1 | tonal | 0.96 | no visible event | |
| 43.56 | 1307 | 1 | click | 1.00 | no visible event | ✓ |
| 44.00 | 1320 | 1 | hit | 1.00 | no visible event | ✓ |
| 45.12 | 1353 | 1 | tonal | 1.00 | no visible event | |
| 45.33 | 1360 | 1 | click | 0.97 | no visible event | ✓ |
| 45.77 | 1373 | 1 | unknown | 1.00 | no visible event | ✓ |
| 46.00 | 1380 | 1 | tonal | 1.00 | no visible event | |
| 46.23 | 1387 | 1 | tonal | 1.00 | no visible event | ✓ |
| 46.67 | 1400 | 1 | unknown | 1.00 | no visible event | ✓ |
| 47.11 | 1413 | 1 | click | 1.00 | no visible event | ✓ |
| 47.51 | 1425 | 1 | hit | 1.00 | no visible event | ✓ |
| 47.72 | 1432 | 1 | click | 1.00 | no visible event | |
| 48.67 | 1460 | 1 | click | 1.00 | no visible event |
Loudness of the reference (measured): −22.1 LUFS integrated, true peak −11.2 dBTP, range 3.1 LU. That is how this film was mastered, not a target.
Loudness targets for your version (common practice per platform; the brief overrides them):
| where it plays | integrated | true peak max | note |
|---|---|---|---|
| YouTube (this reference) | −14 LUFS | −1 dBTP | louder masters are turned down; quieter ones are not turned up |
| X, LinkedIn | −14 LUFS | −1 dBTP | no published normalisation: the file plays as mastered |
| TikTok, Reels, Shorts | −14 to −10 LUFS | −1 dBTP | played on phone speakers; short-form is often mastered hotter |
| Website or app (autoplay) | −16 LUFS | −1 dBTP | often starts muted; keep the unmuted level comfortable |
| Broadcast (EBU R128) | −23 LUFS | −1 dBTP | only when the brief asks for it |
Original music bed (inferred)
Write an original 136 BPM G-minor bed in 4/4: muted kick, round sub, dry rim, soft arpeggiated synth and short glass-pluck accents. Beat 1 uses a thin rising texture; beats 2–5 use a stable pulse with fewer upper notes during UI reading; beat 6 removes the kick and resolves on a sustained minor-to-open chord. Place a whoosh at f6, a hit at f20, restrained tonal accents at f100 and f580, and end punctuation near f1320 and f1425. Use the tracked JSON beats for accents; its bar start is uncertain. Music-model prompt: “Original minimal G-minor electronic product-demo bed, 136 BPM, clean soft pulse, sparse glass plucks, airy final resolve, no vocals or borrowed melody.” Master the new YouTube upload near −14 LUFS, true peak ≤ −1 dBTP.
Plate 09
Packaging
- Title (inferred): “Introducing {{FEATURE}} for {{PRODUCT}}” — direct announcement, no mystery headline.
- Cover / thumbnail (inferred): One clean UI card centred on
{{BRAND_BG}}, one short benefit line and a small{{LOGO}}. - Post copy (inferred): One sentence for the new capability, two concrete uses, availability terms, then
{{CTA_URL}}. - Audience signals: 23,949 views and 108 comments at capture; comments include confusion about what the feature does, so make the new version's first line explicit.
Plate 10
Script template
(inferred) Music-led on-screen copy only. Keep each chapter line under seven words; UI labels state real actions and outcomes.
[beat 1 · shot 1] Your {{PRODUCT}}, now with {{FEATURE}}.
[beat 2 · shot 1] Claim it. Keep your information current.
[beat 3 · shot 1] Add your real catalog.
[beat 4 · shot 1] Answer questions from approved information.
[beat 5 · shot 1] See how people discover you.
[beat 6 · shot 1] Get started with {{PRODUCT}}. {{LOGO}} {{CTA_URL}}
Plate 11
Production plan
Two routes (inferred)
Minimum: Claude alone. Remotion with HTML/CSS and CSS 3D builds the typography, cards, chart, cursor, shallow edge props and an original synthesized music bed at 1920×1080/30 fps. Simplify the edge objects into flat illustrated silhouettes and use only real {{UI_ASSETS}} or code-built UI. It gives up the reference's polished prop textures and some interface detail; the user supplies authentic product data and imagery.
Full: with the toolkit. Design authentic screens and product cards in Figma or code; export 1920×1080 PNGs/PNGs with alpha and a JSON manifest of real labels. Build the animation and camera/element keyframes in Remotion at 1920×1080/30 fps. Create original music and tiny UI sounds in a DAW as 48 kHz WAV with a beat/click track; assemble and master in Resolve or FFmpeg to 1920×1080/30 fps H.264. Handoffs: design → Remotion, alpha PNG sequence plus data JSON; Remotion → edit, ProRes or lossless 1920×1080/30 fps render; DAW → edit, 48 kHz stereo WAV and click track; edit → H.264 MP4.
Asset honesty: Use real product captures or your own UI rebuilt in code; where no image exists, a type tile says so. Never generate product UI, fake screenshots, invented results, reviews or metrics. In beat 5, either use measured analytics from {{PRODUCT}} or label the chart as illustrative.
Plate 12
Prompts
Script prompt
Write six concise on-screen lines for {{PRODUCT}} and {{FEATURE}} using the beat functions in plate 4. Every claim must map to a real screen or supplied {{UI_ASSETS}}. Use sentence case; keep chapter lines under seven words. Close with {{LOGO}} and {{CTA_URL}} as plain text.
Music prompt
Original 136 BPM G-minor minimal electronic bed for a bright UI announcement: soft steady pulse, restrained sub, short glass-pluck accents on reveals, sparse under detailed screens, airy resolved ending; no vocals or borrowed motif.
Plate 13
Edit recipe
How it's built (inferred)
Signature move: title becomes proof. The hook first establishes a fixed central type plane, while small props travel along the perimeter. From f2–12 slide props up about 8.5% frame height, then from f12–52 scale the central moving region to about 0.52 with a slight left/up drift (plate 5). The repeated name in the middle works as a content swap inside a fixed clipping window; make the outgoing text opacity 0 while the incoming word rises 12 px over 8–10 f. At f100–108, move the active element right around 3.4% frame width; after that, reveal a real business card, keeping the white field. Use cubic-bezier(.22,1,.36,1) for the type and card settling; use cubic-bezier(.42,0,.8,1) for the first upward acceleration. The camera itself stays level: plate 5's static? reading is low confidence, and the sheets show element motion, not a dolly.
The five hand-offs. At f200, replace the business card with a claimed-state card on that frame: 0-f hard cut, then build the badge/content within the new phase. At f245, hard cut from card to title; the word build following that cut is element animation. At f682, hard cut from the chapter line to the agent setup UI; reveal text and mobile panel after the cut. At f972, hard cut from product page to the analytics chapter; type completes over subsequent frames. At f1240, hard cut from analytics UI to the call-to-action title. Keep these distinct from the soft changes inside phases. No “no visible cut” boundaries are measured.
Other element methods. For beat 2, move a vertical stack of form cards under a stationary viewport; keyframe the stack y position, not the camera. For beat 3, stage upload tile → table → storefront as content replacements; use real products and keep their cards within one mask. For beat 4, anchor the phone panel at a consistent x position, enlarge it with transform scale and swap the screen content only at a logical tap. For beat 5, build the metric row first, then reveal chart arc segments and the edit panel; the timeline tiles show a title-to-metrics change around 34.2 s and chart build around 36.5 s. Use opacity/clip masks for these, not additional cuts. At the end, let the central imperative yield to the wordmark and coloured dot; no fade to black.
Cadence and blur. Animate on ones at 30 fps; plate 5 reports about 28.2 unique frames per second while moving. Use 180° shutter-equivalent blur (roughly 1/60 s) only on fast prop edges and the cursor; leave text and UI sharp. Export moving props with 2× supersampling to avoid thin-outline shimmer. The aggregate 23.6-s element path in plate 5 spans multiple unrelated objects, so do not apply its −89.1° roll or ×0.24 total scale to one card.
Build order
- Lay out beats 1–6 as one Remotion composition with the five hard-cut frames from plate 5; keep each phase's internal layers separate.
- Implement the type hierarchy and boxes in plate 6; replace all reference-specific words and imagery with the declared slots.
- Build the signature hook motion and the five hand-offs, then the UI content swaps and chart animation.
- Add the original bed to the tracked beat and cue grid in plate 8; master and render.
Plate 14
QA checklist
- Five hard cuts land at f200, f245, f682, f972 and f1240, within ±1 f; all other changes are inside shot 1.
- Essential text is ≥38 px cap at 1920×1080; in vertical or square exports it is ≥22 px cap.
- Hook's measured moving-region segment reaches about ×0.52 by f52, without scaling the whole frame.
- First whoosh and title hit are within ±1 f of f6 and f20; other intentional accents align to the tracked beat/cue frames.
- YouTube master is near −14 LUFS integrated with true peak ≤−1 dBTP.
- No reference brand colour, font, mascot, logo, slogan, or music survives; all shown UI/results are authentic or plainly illustrative.
- The frame remains high-key through the ending; there is no fade to black.
Plate 15
Make it yours
Brand slots
| Slot | Replaces in the reference | Notes |
|---|---|---|
{{PRODUCT}} | Business name and subject | Must match the real demo |
{{BRAND_FONT}} | Reference sans | Use your own family |
{{BRAND_BG}} | Near-white field | Keep enough text contrast |
{{BRAND_ACCENTS}} | Blue actions and colourful edge props | Your own accent system |
{{LOGO}} | Final wordmark | Supplied vector preferred |
{{CTA_URL}} | Business Hub destination | Display as plain text |
{{FEATURE}} | Business claim/agent capability | One verifiable capability |
{{UI_ASSETS}} | Toy-store screens and product photos | Real captures or code-built interface |
What this film avoids (keep these out too):
- No voiceover or talking head.
- No stock footage, fake screenshots or invented customer results.
- No feature-list montage or heavy UI chrome around the white field.
- No QR code, store badges or black closing fade.
Brand-distance warning (inferred): A white field, blue verified badge, colourful retail trinkets and the original product wording together will read as the reference brand. Change the badge shape and edge-object visual language first, then apply your own type and palette.
Adapt the aspect (inferred):
- 9:16: Re-lay out; the measured centre crop keeps only 5 of 27 boxes. Stack UI cards under the title, move edge props inside the safe middle 80%, keep essential type ≥22 px cap, and leave top/bottom platform controls clear.
- 1:1: Centre crop, then move the three lost measured boxes inward. Fit the dashboard as two stacked cards, keep essential type ≥22 px cap, and protect a 10% edge margin for feed overlays.
Adapt the length (inferred):
- 15 s: Keep the premise, one authentic action/proof sequence, and beat 6; drop the optional detailed form and analytics tours.
- 30 s: Keep all six beat functions but compress the UI dwell and show one example per action.
- 60 s: Add real proof within beats 3–5, such as one extra product, one customer question and one interpretable analytics view; keep the opening and end-card timing crisp.
The creator's signature (don't copy): Its wording, brand marks, exact prop set, accent palette and music. The transferable format (do copy): Declarative hook → genuine product steps → concrete outcome → clean text call to action, paced by a restrained beat grid.
Source: https://www.youtube.com/watch?v=Z9k5oI3Paoc · measured 2026-11-11 · (measured) = read off the film · (inferred) = construction reading
For agents: Aspect conversions · Length cuts
The measured data and build specs agents work from. Over MCP they get it all, with the full timeline as JSON.
Aspect conversions (computed from the measured boxes)
Aspect conversions (computed from the measured boxes): a centre crop of the reference frame, checked against its text boxes and moving elements.
| to | deliver at | a centre crop keeps | measured boxes kept | boxes a crop loses | phone minimum | so |
|---|---|---|---|---|---|---|
| 9:16 | 1080×1920 | middle 31.6% of the width (cut 34.2% each side) | 5 of 28 | "now in ChatGPT", "NOVA ONE HEADPHONES", "HARBOR WEST SUPPLY CO", "Toy Shop" … | 22 px cap | re-lay out: a crop loses too much |
| 1:1 | 1080×1080 | middle 56.2% of the width (cut 21.9% each side) | 25 of 28 | "Add products to ChatGPT", "Answer customer questio…", "Top discovery topics" | 22 px cap | centre crop, then move the lost boxes inside |
Length cuts (computed from the scales flags)
Length cuts (computed from the scales flags): frames per shot at 30 fps; fixed shots keep their length, optional shots go first when the cut is shorter, the scaling shots share what is left. Snap each cut to the nearest beat when plate 8 has a steady pulse.
| # | flag | reference (f @30) | 15 s cut | 30 s cut | 60 s cut |
|---|---|---|---|---|---|
| 1 | scales | 1458 | 450 | 900 | 1800 |
| total | 1458 f · 48.6 s | 450 f · 15.0 s | 900 f · 30.0 s | 1800 f · 60.0 s |
15 s: scaling shots shrink to ×0.31, too fast: drop a beat instead.
Recipe
One prompt that makes a film like this with Claude. Use the method; make the footage, words and brand your own.
Make a 48.7 s 16:9 film for {{PRODUCT}} in the format of yt-Z9k5oI3Paoc (white-canvas product walkthrough). Brand: font {{BRAND_FONT}}, background {{BRAND_BG}}, accents {{BRAND_ACCENTS}}, logo {{LOGO}}, call to action {{CTA_URL}}. Feature: {{FEATURE}}. Authentic product screens and images: {{UI_ASSETS}}. Load get_blueprint("yt-Z9k5oI3Paoc", format: "json") first and follow its timing and motion data. Make this music-led and wordless. Beat 1 states the new capability with perimeter props and cycling examples; beat 2 shows claiming and updating; beat 3 shows adding actual products; beat 4 shows a genuine customer answer; beat 5 shows real or clearly illustrative analytics; beat 6 gives one imperative. Build the signature hook with a fixed central text plane and upward-moving perimeter elements, then scale the active region with eased transforms and sharp type. Keep chapter text and essential UI labels readable at 1080p. Render the full route in Remotion from code-built UI and supplied captures, with original music produced as 48 kHz WAV; the minimum route uses Remotion with CSS 3D and Claude-synthesized music. Follow the tracked 136 BPM grid; master YouTube audio near −14 LUFS, true peak at most −1 dBTP. Output 1920×1080, 30 fps H.264 MP4. End on {{LOGO}} plus {{CTA_URL}} as text, with no QR code or store badges. Never generate product UI, fake screenshots or fake results; never copy the reference's colours, font, logo, mascot, slogans or music.Films like this
Tagged product launchui animationkinetic typebusiness softwaresilent viewingcustomer chatdashboardwhite background






