One prompt controls an iPhone ride app
A silent, continuous split-screen demo shows an AI agent navigating a real phone app up to a priced ride confirmation.
Field notes / why it works
The task is typed in the first 2.75 seconds while the phone and terminal stay visible throughout, letting viewers compare the agent's actions with the app's response. The app opens around 8–10 seconds, a destination choice appears around 13.6 seconds, and a $41.40 confirmation at about 22 seconds gives the demo a concrete payoff; the clip does not establish a completed booking.
Plate 01 / The format in one breath
A silent, uninterrupted desktop recording puts a phone on the left and an agent terminal on the right. A plain-English request appears immediately; the viewer watches the agent operate an app, resolve a choice, and reach a consequential confirmation prompt. The film sells capability through visible UI evidence rather than a narrator or title cards.
Why it works
- The first 2.75 seconds build a specific, testable request inside the terminal while the phone is already visible; the viewer can anticipate what should happen next.
- The phone and terminal stay side by side for all 30.4 seconds, so the app changes around 8–10, 18.6, and 22 seconds can be compared with the agent's actions.
- The exact destination choice around 13.6 seconds and the $41.40 fare confirmation around 22 seconds give the demonstration concrete stakes. The last visible frames show a request to proceed and continued processing, not proof of a completed booking.
Plate 02 / Format card
| Platform · aspect · length | X · 1106×720, 553:360 (~1.54:1) · 30.4 s · 30 fps |
| Pace | One continuous composite shot; 0 cuts/min; 8 measured internal picture changes, ~15.8 changes/min; 30.4 s shot |
| Script | No spoken script; one typed natural-language request, terminal status text, choices, and app UI |
| Voice | None; no person appears |
| On screen | ~100% desktop screen recording: phone mirror left, terminal right; no b-roll |
| Captions | None added; native app and terminal text carry the story |
| Sound | Effectively silent: measured −70 LUFS; no discernible music or effects; true peak unavailable |
| Structure | Request → tool starts → device navigation → choice → confirmation → response/processing; no in-video 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- 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