A dinner problem becomes a browser context reveal
A cinematic ensemble dinner scene turns a growing list of constraints into a quiet, tab-aware browser demonstration.
Field notes / why it works
The film earns the demo with a concrete human problem: ingredient confusion starts by 0:07, then dietary and contextual constraints pile up from 0:19 to 0:34. A quieter turn at 0:38 lets the tab-aware answer at 0:42–0:56 resolve those exact details; the strongest replay peak at 0:43.7–0:44.5 falls on the source-selection reveal. The tagline arrives at 0:58.5, after proof, while a warm coda returns the technology to people.
Plate 01 / The format in one breath
A brisk, warmly lit ensemble scene turns a mundane request into an impossible-looking memory task. At 0:34–0:40, a character realizes the product already has the relevant context; a quiet, legible UI demonstration then resolves the exact constraints the characters named. The explicit product promise waits until 0:58.5, after the result.
Why it works
- The first cut arrives at 0:00.25 and the opening 10 s has six cuts: rapid food and character details establish a real situation before any sales language.
- The dialogue at 0:20–0:34 stacks recipe, allergy, timing and family-theme constraints; the UI answer at 0:49–0:56 can therefore prove more than a generic chatbot response.
- The 0:43.7–0:44.5 replay peak lands on selecting existing tabs and the first context card, precisely where the promised capability becomes concrete. The second-highest peak is 1:06.3–1:07.1 in the late coda.
Plate 02 / Format card
| Platform · aspect · length | YouTube · 16:9, 1280×720, 24 fps · 77.8 s |
| Pace | 45 detected shots; 33.9–40.9 cuts/min allowing uncertain fast changes; 42.4 measured picture changes/min; mean shot 1.73 s, median 1.25 s. Last detected shot runs 11.08 s. |
| Script | Automated transcript: 108 words, 89 wpm across runtime, 209 wpm during detected speech; 21 short utterances/sentences. Some food words are low confidence. |
| Voice | Naturalistic ensemble dialogue: a composed host/narrator, quick helpers, then a quieter pleased coda. No omniscient ad voice. |
| On screen | Roughly 55% cinematic live action, 28% photographed/filmed interface, 17% black title or text cards (estimated from sheet frames and timings). |
| Captions | No burned dialogue subtitles; only diegetic UI text and two spare black intertitles. |
| Sound | Continuous tonal bed is detected; kitchen/dialogue ambience and cut accents are plausible but effect identification is uncertain. Measured −17.8 LUFS integrated, −2.7 dBFS true peak. |
| Structure | 7 beats: situation → escalating constraints → chat as workaround → context realization → tab-aware demo → result/tagline → comic human coda. |
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- Plate 03 / The hook, frame by frame (first 3 seconds)Pro
- Plate 04 / Structure (beat sheet)Pro
- Plate 05 / Shot grammarPro
- Plate 06 / Visual style guidePro
- Plate 07 / Voice and deliveryPro
- Plate 08 / Sound designPro
- Plate 09 / PackagingPro
- Plate 10 / Script templatePro
- Plate 11 / Production planPro
- Plate 12 / PromptsPro
- Plate 13 / Edit recipePro
- Plate 14 / QA checklistPro
- Plate 15 / Make it yoursPro
- Recipe / runnable promptPro