Spark's scattered-to-searchable product film

A small typographic claim turns into scattered source imagery, then a sequence of actual product tasks proves that one place can capture, retrieve and assemble the viewer's material. Narration names the problem and each action; UI motion supplies the evidence.

Why it works

• The word build in shot 1 lasts until the dock appears after its internal cut at frame 77, then the first shot boundary waits until frame 123.
• The retrieval beat gets shots 17–19, including the search-to-gallery hand-off at frame 1002, rather than a single generic claim.
• Shot 26 begins at frame 1781 and gives the final brand message a measured 123-frame allocation without a closing fade.

Platform · aspect · length	X · 16:9 · 63.50 s · source 1280×720 · deliver 1920×1080 · 30 fps
Pace	26 measured shots · 17 boundary hard cuts plus internal cuts · median shot 2.17 s
Script	Narration-led · 149 words · 143 wpm overall, 176 wpm during speech; cuts follow phrases, not a beat grid
Voice	Conversational product guide, bright and precise; no visible speaker
On screen	UI capture, dimensional card and folder composites, typographic bridges, brand end card
Type	Eight measured cap-height levels, 21–266 px; enlarge essential small UI words to the 38 px phone minimum
Sound	Continuous tonal bed; no steady pulse; cue specific accents; reference −15.5 LUFS
Structure	Eight beats; retrieval is the extended payoff; the direct call lands in shot 26

Tools you'll need: screen recordings of the product · Remotion for motion graphics · ElevenLabs or a recorded voice · Remotion or any video editor · Claude alone: partly