# Type a four-line staircase title, cut the background under it, then pop the lines off bottom-up

Manual DaVinci Resolve / Fusion construction guide generated from the recipe. It is not a tested .drp.
The accompanying Remotion project (recipe.json) is the executable definition; the numbers below are copied from it.
Neither construction establishes source fidelity: this is an original vector approximation of an inspected editing moment.

## Source and clock
Reference: OpenAI's "Plugins & Skills" (5QPBVrAAdBk), published 2026-08-28.
Source interval [18, 112) on the 24000/1001 fps master: 94 frames, 3.920583 s. Local frame 0 = source frame 18.
Create a 1920x1080 progressive timeline at 24000/1001 fps and a Fusion composition exactly 94 frames long. Keep integer frames; do not round the frame rate.
Background field: #b6bfc4. Export silent H.264 yuv420p to match the intended playback container.

## Observation
Over a bright top-down room shot the first word types on at local ~5-12 in the top-left, the second word types in one step down and right (~13-18), the third word (~22-30) and a smaller fourth line (~34-48) continue the staircase toward the bottom-right. At ~48 the footage hard-cuts to a darker laptop close-up while all four lines stay in place. From ~78 the lines pop off from the bottom up: line four at ~80, line three at ~82, line two at ~86, line one at ~88.

## Mechanism
Staircase title build: sequential character type-ons stepping diagonally down the frame, held across a hard picture cut, then removed bottom-up in quick hard pops.

## Phases (local frames)
- 5: Line one types on
- 13: Line two types on
- 22: Line three types on
- 34: Small line four types on
- 48: Hard cut to darker plate, title holds
- 80: Lines pop off bottom-up

## Layer inventory (absolute 1920x1080 top-left pixels, back to front)
- plate-b [rect] at (0, 0) size 1920x1080 fill #7c8387 visible [48, 94)
- sofa [rect] at (180, 120) size 1100x320 radius 30 fill #1f4a58 visible [0, 48)
- lid [rect] at (-100, 300) size 1500x900 radius 40 fill #5a5f63 visible [48, 94)
- l1 [text] at (260, 110) text "Level up" 84px weight 600 #ffffff reveal chars frames 5-12 visible [0, 88)
- l2 [text] at (700, 320) text "your assistant" 84px weight 600 #ffffff reveal chars frames 13-18 visible [0, 86)
- l3 [text] at (1150, 540) text "with" 84px weight 600 #ffffff reveal chars frames 22-30 visible [0, 82)
- l4 [text] at (1180, 790) text "Add-ons & Recipes" 44px weight 500 #ffffff reveal chars frames 34-48 visible [0, 80)

## Construction steps
1. Timeline 1920x1080 at 23.976 fps; Fusion comp 94 frames. Local 0 = master frame 18.
2. Two picture plates: bright plate (#b6bfc4 room) frames 0-47, darker plate (#7c8387 laptop close-up) from frame 48 as a hard cut on the same track; do not dissolve.
3. Four Text+ nodes, white, weight 600: 84 px at (260,110), (700,320), (1150,540); 44 px at (1180,790). Each steps down-right to form a staircase.
4. Write On (character mode, no caret) frames 5-12, 13-18, 22-30, 34-48 respectively; each line starts as the previous one finishes.
5. Exit: cut each text node off bottom-up at frames 80, 82, 86, 88 (visibility off, no fade).
6. Easy mistake: the title must not react to the cut at 48; only the background changes, the text stays pixel-locked.

## Content variants
Default copy:
- l1: Level up
- l2: your assistant
- l3: with
- l4: Add-ons & Recipes
Adaptation changes:
- l1: Do more
- l2: every week
- l3: using
- l4: Shortcuts & Routines

## Limits
Easing names map to Fusion spline shapes: linear = linear, in/out/inOut = ease curves on the same frames, hold = step, spring/overshoot = slight overshoot past the target before settling.
Text reveal frames give the first and last unit; distribute characters/words evenly between them. Cursor click rings last 14 frames.
Vector illustration is an approximation of layout and motion, not the product UI or recovered production assets.
