Skip to content
refs
REF-2796Source onlineBlueprint rev 01

Voice Agent Launch in a Terminal

A sparse voice-agent announcement turns into a live terminal exchange that proves interruption and tool use.

Maker @vercelFor VercelOriginal post ↗
Cut timeline · 1 shots00:58
00:0000:1000:2000:3000:4000:50

Field notes / why it works

A tiny animated icon and typed capability claim establish the promise by 0:08 without a cut. The terminal then shows a real conversational test: a response around 0:22, interruption around 0:29, and a tool-backed answer around 0:45. The only major composition reset at 0:51.6 gives the dense proof a clean closing claim.

01

Plate 01 / The format in one breath

A black-screen product announcement becomes a live, text-visible voice-agent test. It makes a capability claim in the first 8 seconds, leaves space for a real interaction, and closes with a compact product thesis.

Why it works

  • The promise is legible before the demo: small product label at 0:00.25, large capability headline typed from 0:02.2, complete by about 0:07.7.
  • The viewer sees the test conditions and exchange in a terminal from about 0:11–0:50; the agent's interruption and tool use are demonstrated around 0:30–0:49.
  • The 51.6 s transition returns to the animated mascot and a four-word closing claim, converting a dense demo into a memorable end frame.
02

Plate 02 / Format card

Platform · aspect · lengthX · 16:9, 1280×720 · 57.7 s at 30 fps
Pace1 continuous shot; 0 hard cuts/min; one major composition change at 51.6 s. Typing and mascot animation create smaller changes throughout.
Script106 measured spoken words · 128 wpm across the full runtime, 234 wpm while talking · 13 short sentences, 2 questions
VoiceA concise medium-register announcer, then a conversational human tester and responsive agent voice (speaker roles inferred from dialogue)
On screenOpening and closing motion graphic, about 14 s total; terminal-style demo, about 44 s; no faces
CaptionsNo conventional subtitles. Tiny monospaced terminal lines act as the visual transcript.
SoundDry voices and digital silence; no measured music bed or clear cut effects · −27.9 LUFS integrated, −7.7 dBFS true peak in source
StructureCapability statement → terminal setup → live exchange → tool-result proof → black end card; no spoken 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