Your evening, handled.
Cinematic key frames, brand-accurate UI stills at native screen resolution, a designed end card, and a frame-accurate handoff so an editor can composite the final 30 seconds.
Character lock
Locked identity fed into every clip frame as a visual anchor: early-30s, light-olive skin, sleek low bun, camel wool coat, cream silk blouse, black leather tote, natural makeup, gold huggie earrings.
Approve this sheet before generating Phase 3 frames — they reference it directly.
Generation prompt
Character reference sheet, vertical 2:3 portrait layout, of one single woman. Woman in her early 30s, warm light-olive skin, dark chestnut hair in a low sleek bun with soft face-framing strands, almond-shaped dark brown eyes, natural 'no-makeup' makeup with a soft rose lip, small gold huggie earrings. Wearing a tailored camel wool coat over a cream silk blouse, carrying a structured black leather tote bag. Elegant, composed, quietly tired professional energy. Layout: a large front-facing medium close-up at the top, and below it a row of three smaller panels: three-quarter profile turned left, clean side profile, and a seated back-seat-of-a-car angle lit by passing city lights. Neutral warm porcelain studio background, soft even key light so face, hair and coat are perfectly legible. Consistent identical face and outfit in every panel. Cinematic live-action photography, shot on ARRI Alexa with 35mm anamorphic lens, shallow depth of field, natural motivated lighting, filmic color grade with warm amber highlights and cool teal-charcoal shadows, subtle film grain, premium fashion-commercial look. Photorealistic, no illustration, no text, no logos.
UI stills & end card
Designed in the Marova system (Cormorant Garamond + Inter). Phone stills at 1080×1920, tablet at 1536×2048 — render to PNG for corner-pinning. Maya 6:15 · Lena 7:20 · $165 stays identical everywhere.
Live-action clips
Key frame per clip, anchored to the reference sheet, plus a ready-to-paste video prompt. Each clip can be rendered to an 8-second 9:16 MP4 with Kling v3 Pro (image-to-video, audio off) straight from its key frame. Phones and tablets stay blank in every clip.
Office exit + walk to car
9:16 vertical, 8 seconds, 24fps. Smooth gimbal tracking shot moving backward ahead of a woman in her early 30s (match reference image: sleek low bun, camel coat, black leather tote) as she exits a Midtown Manhattan glass office lobby at evening rush hour and walks toward a black sedan at the curb. Yellow taxis and pedestrians pass in soft motion blur, wet street reflections, blue-hour sky. She glances at her phone once — screen stays dark. Cinematic anamorphic look, shallow depth of field, warm/teal filmic grade, subtle grain. No on-screen text, no logos, no readable UI.
Back seat, she opens phone
9:16 vertical, 8 seconds, 24fps. Interior black sedan, night, Manhattan. Slow push-in from over the shoulder to a three-quarter profile of the same woman (match reference). She lifts her phone with both hands; the screen is blank with only a soft warm glow on her face. Passing city lights and taxi headlights streak as bokeh across the window. Her expression shifts from tired to focused. Locked-off camera with gentle handheld micro-movement. Cinematic filmic grade, subtle grain. No readable UI on the phone.
Match cut — Aura Med Spa front desk
9:16 vertical, 8 seconds, 24fps. Warm minimalist med spa reception at evening: porcelain plaster walls, pale oak desk, blush linen, eucalyptus, soft cove light. A calm manager in her 40s with silver-streaked hair picks up a tablet from the desk, its screen a soft blank glow, studies it, taps once, and gives a small satisfied nod. Slow dolly-in from medium to medium-close. Shallow focus, cinematic grade, subtle grain. No readable UI on the tablet, no text.
Return to car — relief
9:16 vertical, 8 seconds, 24fps. Back in the black sedan at night. Same woman (match reference) lowers her phone to her lap, screen dark, and turns to look out the window, exhaling with a soft relieved smile. Warm amber storefront light glides across her face and the glass. Slow, gentle rack focus from phone to her eyes. Close-up, seat-beside-her angle. Cinematic filmic grade, subtle grain. No text.
Warm minimalist vs. SoHo loft-modern
Same manager, same blank tablet, same dolly-in — only the Aura interior changes. Use the picker above to lock one; the S3 hero, Clip 3 and the rough cut follow.
Porcelain limewash, pale oak, blush linen, eucalyptus. Reads closest to the brand palette; end card feels continuous.
Concrete, black steel, greenery, brass, one burgundy velvet bench. More Manhattan-identifiable; contrast against porcelain UI is stronger.
Variant video prompt
9:16 vertical, 8 seconds, 24fps. SoHo loft-modern med spa reception at evening: polished concrete, black cast-iron columns, black steel windows with city glow, monolithic concrete desk, trailing greenery, brass pendants, a burgundy velvet bench. A calm manager in her 40s with silver-streaked hair and a black uniform picks up a tablet from the desk, its screen a soft blank glow, studies it, taps once, and gives a small satisfied nod. Slow dolly-in from medium to medium-close. Shallow focus, cinematic grade, subtle grain. No readable UI on the tablet, no text.
One-click preview assembly
Joins the four clips and the end card on the handoff timing map, burns in the three headline titles at their cue points, and drops the S1–S6 stills roughly onto the blank phone and tablet so the UI timing reads at a glance. Where a Kling MP4 isn't rendered yet, the key frame is animated with a slow push-in. Spa style: warm.
Editor handoff
| Time | Clip | Picture | Overlay | VO | Sound |
|---|---|---|---|---|---|
| 0:00–0:08 | Clip 1 | Office exit, walk to black car. Hold full 8s, trim 0.5s head. | None. Optional lower-third at 0:03: 'Manhattan · 6:02 PM' (Inter 500, 40px, porcelain, 70% opacity, bottom-left safe zone). | VO line 1 | City ambience, heels on pavement, distant cab horn. |
| 0:08–0:16 | Clip 2 | Back seat. Corner-pin S1 at 0:09.5 → S2 at 0:11.5 → S3 at 0:14. Cross-dissolve 8 frames between stills. | Headline 0:12–0:15: 'One request.' (Cormorant 600, 96px, porcelain, upper-third centered, 2% letter-spacing). | VO lines 2–3 | Car interior hush, soft UI chime on S2 parse (0:11.8), gentle 'send' whoosh on S3 (0:15). |
| 0:16–0:24 | Clip 3 | Match cut on phone glow → tablet glow. Corner-pin S4 at 0:17 → S5 at 0:20.5. | Headline 0:19–0:22: 'Every service, planned.' (same style). | VO line 4 | Spa ambience (soft water, low pad), notification tone on S4 (0:17.3), plan-sent chime on S5 (0:23). |
| 0:24–0:27 | Clip 4 | Return to car. Corner-pin S6 (confirmed) 0:24–0:26, then she lowers phone. Use clip4 0–3s. | None over face. | VO line 5 | Confirmation chime (0:24.5), music resolves to warm major chord. |
| 0:27–0:30 | End card | 1s cross-dissolve from clip4 into end card. Hold to 0:30. | Burgundy sparkle particle (AE CC Particle World or Trapcode) drifting top-right of wordmark, 8% opacity, 0:27.5–0:30. | VO line 6 | Music tail, single soft sparkle transient at 0:27.8, room tone out. |
- 0:01“It's 6 o'clock. You've done the work. You just want the evening to work too.”
- 0:09“Booking two things at two places… used to mean seven apps and a phone call.”
- 0:12“Now it's one request.”
- 0:17“Marova sends it straight to the spa — and they plan it around you.”
- 0:24“Confirmed. That's it.”
- 0:27“marova. Your evening, handled. Join the waitlist — launching first in San Diego.”
- tool
- ElevenLabs (Multilingual v2 or Turbo v2.5)
- voice
- Female, late 20s–30s, warm, low-mid register, unhurried. Suggested presets: 'Charlotte' or 'Matilda'. Stability 0.55, Similarity 0.75, Style 0.15, speaker boost on.
- delivery
- Intimate, near-whisper confidence — talking to a friend in the back of a car. No 'ad voice'. Slight smile on line 5.
- processing
- High-pass 90Hz, gentle de-ess, 2–3 dB compression, tiny room reverb (0.4s) to sit inside the car.
- 0:00Music in — soft felt piano + warm sub pad, 78 BPM, key of D♭ major, minimal.
- 0:00–0:08Manhattan exterior bed: traffic, heels, cab horn far right.
- 0:08Car door closes (soft, expensive). Interior hush, low tire rumble.
- 0:11.8Chime A — parse complete (two-note mint bell, 400ms).
- 0:15.0Whoosh — request sent (airy, 300ms).
- 0:16Match-cut: interior rumble → spa bed (water trickle, soft pad). Low-pass sweep 200ms.
- 0:17.3Chime B — new request notification (single soft tone).
- 0:23.0Chime C — plan sent (ascending two notes).
- 0:24.5Chime D — confirmed (warm triad, longer 700ms tail). Music resolves.
- 0:27.8Sparkle transient (glassy, very quiet).
- 0:30Music tail out, 400ms fade.
- lut
- Start from Kodak 2383 print emulation at 60% or FilmConvert 'Kodak Vision3 250D'.
- look
- Warm amber highlights (skin, cab lights), cool teal-charcoal shadows in clips 1–2. Clips 3–4 shift warmer and softer: lift blacks to ~4%, reduce saturation −8, add blush tint to midtones (hue toward #F6E1E4).
- skin
- Protect skin: hue-vs-sat keep skin at 100%, slight highlight roll-off. Halation subtle on car window lights.
- grain
- Fine 35mm grain, 12–15% intensity, consistent across all clips and the end card.
- end card
- No grade on the end card — porcelain must read as exact #FBF8F7. Apply grain only. QR (marova.com/waitlist) must stay unblurred and ≥ 180px on the 1080-wide master.
- Stills are delivered at native screen resolution (phone 1080×1920, tablet 1536×2048). Corner-pin onto the blank device screen with Mocha/planar tracking.
- Add a 2% screen reflection layer (soft white gradient, 'screen' blend, 6–10% opacity) over every still so it sits in the device.
- Feather the still edge by 2px and add a 1px darker inner bevel to sell the glass depth.
- Motion-blur the still along with the tracked device (enable motion blur on the corner-pin layer).
- Keep all prices, names and times identical to the stills: Maya 6:15 PM Hydrating Facial · Lena 7:20 PM Brow Shaping · total $165.
- Text overlays live in the 9:16 safe area (avoid top 250px and bottom 320px for platform UI).
- format
- 1080×1920 (9:16), 24fps, ProRes 422 HQ master + H.264 delivery
- duration
- 30.00s (4×8s raw = 32s; trim transitions as noted)
- audio
- Stereo 48kHz, −14 LUFS integrated for social
- fonts
- Cormorant Garamond (editorial headings, 500/600), Inter (interface, 400/500/600)
- palette
- Porcelain #FBF8F7 · Blush #F6E1E4 · Burgundy #6D2433 · Mint #D9F1E1 · Lilac #EAE4F8 · Charcoal #1F1F1F