Skip to content

SCRIPT.md

Latest commit

 

History

History
243 lines (161 loc) · 5.66 KB

SCRIPT.md

File metadata and controls

243 lines (161 loc) · 5.66 KB

SCRIPT — website-to-hyperframes Launch Video

Duration target: ~61 seconds (Act 3 expanded from 11.2s → 23.5s for full-breath clips + logo-reveal section) Voice: ElevenLabs — Rachel (21m00Tcm4TlvDq8ikWAM) Voice settings: stability 0.35 · similarity 0.75 · style 0.20 · speaker boost: on Model: eleven_multilingual_v2

Voice direction: Curious, playful on Act 1 questions. Confident and slightly smiling on the Act 1 announcement. Rhythmic, punchy, trailer-cadence on the Act 3 capability listing. Calm + warm on the Act 4 CTA. Always feels like a person, never like a reader.

Act 1 — Cold Open (0:00 – 0:15)

Four questions under the 4 typing panels, then the announcement, then silence through YEAHH → click → drop.

Line 1 — Panel 1 (stripe.com)

Time: 1.66 – 3.9s Delivery: Upward lilt at the end — genuine curiosity.

Promo for your app?

Line 2 — Panel 2 (framer.com)

Time: 4.2 – 6.4s Delivery: Slightly brighter than line 1, same curious register.

Brand video?

Line 3 — Panel 3 (vibecode.app)

Time: 6.7 – 8.9s Delivery: Quicker, punchier.

Social ad?

Line 4 — Panel 4 (heygen.com)

Time: 9.2 – 11.4s Delivery: Shortest, slight anticipation — "one more..."

Launch clip?

Line 5 — The Announcement

Time: 11.7 – 13.5s Delivery: Drop into confidence. Period after "Now" creates a beat. "any site" gets gentle emphasis. Ends with a slight smile the listener can hear.

Now — any site becomes video.

Pronunciation cues: "website-to-hyperframes" is never spoken here; we save the product name for Act 4 to give it weight.

Act 2 — Extraction (0:15 – 0:30)

No VO. Music carries the full energy. Only on-screen monospace labels:

  • extracting fonts
  • capturing colors
  • parsing assets
  • detecting libraries
  • composing timeline
  • rendering frames

These fade in/out sync'd to extraction beats — the visual is the story. Narration here would be over-explaining.

Act 3 — Capability Reel (0:30 – 0:53.5)

Eleven clips at 0.8s–3.5s each, variable pacing. Rachel reads each capability label rhythmic, period-stopped, trailer cadence. New in v2: logo-reveal trio (linear / dribbble / tailwind-logo) at the end + vercel world map swapped in + shader/figma removed.

Line 6 — Stripe opener (30.0 – 31.5s · 1.5s)

Product launches.

Line 7 — Tailwindcss 6-scene rapid-fire (31.5 – 35.0s · 3.5s)

Feature showcases.

Line 8 — Framer holo-gradient (35.0 – 38.0s · 3.0s)

Brand reels.

Line 9 — Spotify MUSIC FOR (38.0 – 40.0s · 2.0s)

Music apps.

Line 10 — HeyGen 3D cube (40.0 – 42.0s · 2.0s)

Three D scenes.

Pronunciation cue: 3D → three D

Line 11 — Vercel Edge Network / world map (42.0 – 44.0s · 2.0s)

NEW line in v2 (was "Dev metrics" for the old 10.5× FASTER clip; changed to match world map / global infra visual).

Infrastructure.

Line 12 — Vibecode iPhone (44.0 – 46.0s · 2.0s)

Product demos.

Line 13 — Linear logo reveal (46.0 – 48.0s · 2.0s)

NEW line in v2.

Brand openers.

Line 14 — Dribbble logo reveal (48.0 – 51.0s · 3.0s)

NEW line in v2.

Logo reveals.

Line 15 — Tailwindcss logo SVG draw (51.0 – 52.7s · 1.7s)

NEW line in v2.

Drawn in S V G.

Pronunciation cue: SVG → S V G (three letters, spaces force spell-out)

Line 16 — Custom Clip B typography (52.7 – 53.5s · 0.8s)

Typography.

Dropped from v1

  • Custom shaders. (clip-A shader was removed from Act 3 — visually too abstract for the montage)
  • Portfolios. (dribbble now plays its logo-reveal, not the gallery grid)

Unused VO files (kept on disk for reference / future use): 09-custom-shaders.mp3, 12-portfolios.mp3, 18-team-design.mp3.

Act 4 — CTA + End Card (0:53.5 – 0:61.5)

Three beats, sparse pacing. Room for breath between each — matches Thursday's "HyperFrames. Go make something." Calm, confident. (Shifted later due to Act 3 expansion; same internal structure.)

Line 17 — Product name

Time: 53.5 – 55.0s Delivery: "HyperFrames" as ONE compound word — not Hyper then Frames. Slight stress on the first syllable.

HyperFrames.

Pronunciation cue: hyper-frames (single word, camelCase is typographic only)

Line 18 — The promise

Time: 55.5 – 58.5s Delivery: Two-beat structure. Pause after "URL."

Give your agent a URL. It does the rest.

Pronunciation cue: URL → U R L (three letters)

Line 19 — The sign-off

Time: 59.5 – 61.0s Delivery: Slight smile in voice. Warm, confident, not sales-y.

Go make something.

Totals

Line Duration (approx)
4 questions (Act 1) ~4.8s combined
1 announcement (Act 1) ~1.8s
8 capability labels (Act 3) ~9s combined
3 CTA lines (Act 4) ~5.5s combined
Total voice time ~21s of 50s

The remaining ~29s is music + SFX + visual-only moments (Act 2 entirely, YEAHH through drop, final beat of Act 4). Voice sits above music via ducking (music -20dB during VO, full during gaps).

Files generated from this script

Each line becomes its own .mp3 file at launch-video-2/audio/vo/:

01-promo-for-your-app.mp3
02-brand-video.mp3
03-social-ad.mp3
04-launch-clip.mp3
05-announcement.mp3
06-product-launches.mp3
07-feature-showcases.mp3
08-brand-reels.mp3
09-custom-shaders.mp3
10-product-demos.mp3
11-three-d-scenes.mp3
12-portfolios.mp3
13-kinetic-typography.mp3
14-hyperframes.mp3
15-give-your-agent-a-url.mp3
16-go-make-something.mp3