Skip to content
refs
REF-2091Source onlineBlueprint rev 01

Project Genie: Worlds You Can Play

A widescreen launch film turns text and image inputs into a rapid gallery of interactive worlds, with a hands-on image-to-world proof sequence.

Maker @GoogleDeepMindFor Google DeepMind / Project GenieOriginal post ↗
Cut timeline · 29 shots02:19
00:0000:3001:0001:3002:00

Field notes / why it works

The 0:00–0:35 opening grows a dark gallery of worlds into a huge product reveal before showing the creation UI. From 0:55–1:24, mostly 1–4 second scenes demonstrate strikingly different characters and environments; the highest replay peak at 1:26.2–1:27.6 coincides with hands and a phone capturing a cardboard figure before its generated-world payoff.

01

Plate 01 / The format in one breath

A widescreen, music-led launch film opens with a dark gallery of possible worlds, names the product, then shows a chain of increasingly varied interactive results. Sparse white headline cards make the promise; UI captures, playable scenes and brief live-action inputs supply proof. There is no verified narrator or spoken script.

Why it works

  • The opening withholds the title until about 0:17.6: small windows of contrasting worlds appear against black before the large product-name reveal.
  • The middle turns an abstract claim into repeated inputs and outcomes: an environment prompt around 0:28.5, a mountain world at 0:39, a changed character at 0:52–0:56, and several new worlds by 1:24.
  • The replay peak at 1:26.2–1:27.6 lands on hands arranging a cardboard figure and a phone pointing at it; the film then shows that figure inside a generated tabletop world at 1:29–1:35.
02

Plate 02 / Format card

Platform · aspect · lengthYouTube · 16:9 · 2:18.9 · measured file 640×360 at 23.98 fps
Pace29 measured shots; 28 cuts; 12.1–16.4 possible cuts/min including uncertain changes; mean 4.79 s, median 2.55 s. 11 internal changes yield about 16.8 picture changes/min. The opening is a 35.2 s composite sequence; the fastest stretch is 0:55–1:27.
ScriptNo verified speech; about 8–12 short on-screen lines, usually 2–6 words. The automatic transcript misidentified music as speech.
VoiceNone verified. The product is explained through text, UI and picture.
On screenOpening/end motion graphics; generated interactive world capture; brief product UI; live-action hands, phone and cat. No talking head.
CaptionsNo speech subtitles. White sentence-case feature headlines and large product title, mostly centered.
SoundContinuous tonal music-like bed; roughly 110 BPM (heuristic). Integrated −14.0 LUFS; true peak −0.9 dBFS. Distinct cut effects are not established.
Structure6 beats: possibility gallery → name/promise → prompt-to-world → varied capabilities → image-to-world proof → broad promise and CTA.

Pro / 13 plates + recipe

The first plates are open. The full dissection is Pro.

Get the timed shots, type and sound specs, and the runnable recipe.

Open the full blueprint
  1. Plate 03 / The hook, frame by frame (first 3 seconds)Pro
  2. Plate 04 / Structure (beat sheet)Pro
  3. Plate 05 / Shot grammarPro
  4. Plate 06 / Visual style guidePro
  5. Plate 07 / Voice and deliveryPro
  6. Plate 08 / Sound designPro
  7. Plate 09 / PackagingPro
  8. Plate 10 / Script templatePro
  9. Plate 11 / Production planPro
  10. Plate 12 / PromptsPro
  11. Plate 13 / Edit recipePro
  12. Plate 14 / QA checklistPro
  13. Plate 15 / Make it yoursPro
  14. Recipe / runnable promptPro