# Tokyo After Dark — 15s Smartphone Travel Ad A high-density vertical travel prompt that keeps Shibuya geography, one adult host, the same friend group, food action, and a readable five-shot story inside 15 seconds. - Access: free - Models: Provider-neutral prompt architecture - Formats: 9:16 - Duration: 15 seconds - Last tested: July 26, 2026 - Version: 1.0 ## Production outcome A vertical Tokyo travel memory where the destination remains recognizable even during character and food close-ups. ## Why this is in the library Derived from the production prompt behind FlareClip's published Tokyo After Dark output. The archived record does not name the original provider, so this release is intentionally labeled provider-neutral. ## What this prompt solves Travel prompts often lose the city as soon as the camera moves closer to a person. This version gives Tokyo priority over decoration, assigns separate responsibilities to character and city references, and requires recognizable geographic anchors in at least three shots. ## Complete production prompt Copy the full prompt below, then replace the host identity and location-reference notes only when you have equivalent locked assets. Keep the priority order, time blocks, acceptance checks, and targeted negatives intact on the first test. ```text OUTPUT CONTRACT: Create one complete 15-second native 9:16 premium smartphone travel advertisement with five distinct hard-cut shots. Natural 24–30fps motion, realistic phone exposure and autofocus. Do not generate captions, logos, prices, watermarks or readable foreground text; add exact typography in post. NON-NEGOTIABLE CREATIVE PRIORITIES: 1. Tokyo must be recognizable without titles or audio. 2. Keep one clearly adult fictional host and the same three adult friends in every shot. 3. Preserve the host's exact summer-night wardrobe, hair and accessories. 4. Friends, food and reactions support the city; they never replace the location. REFERENCE INPUT ROLES: Reference A locks the host's identity, hairstyle, proportions and wardrobe only. Reference B locks Shibuya crossing geometry, railway infrastructure, night-light balance and crowd scale only. Neither reference may replace the other, introduce a real-person likeness or supply readable brand text. CHARACTER IDENTITY — STRICT CONTINUITY: One fictional clearly adult 27-year-old East-Asian woman with a refined soft-oval face, warm almond-shaped dark-brown eyes, softly arched brows, a slim straight nose, natural rose lips, realistic luminous skin and long espresso-brown loose waves with airy curtain bangs. She is confident and spontaneous, never over-filtered or influencer-posed. WARDROBE — LOCKED: Warm-ivory asymmetric one-shoulder top, pale-blue tailored denim mini skirt, very lightweight ivory linen shirt worn open, cream strappy low-heel sandals, small brushed-gold hoops, one fine gold necklace and a compact deep-chocolate shoulder bag. Keep every item identical. Fabric, hair and bag react naturally to walking and humid night air. ENVIRONMENT — STRICT LOCATION IDENTITY: One coherent Shibuya night route connecting the multi-directional scramble crossing, station-side vertical light canyon and a nearby narrow yokocho beneath visible commuter railway infrastructure. Retain diagonal crossing geometry, Japanese pedestrian signals and road markings, compact black taxis with white roof lights, elevated silver train, vending-machine glow, red lanterns, noren curtains, steam, stainless counters, parked bicycles and slightly wet pavement. Never substitute a generic Asian market, cyberpunk city, Seoul, Bangkok or Hong Kong. CAMERA, LIGHT AND COLOR: Recent flagship smartphone. Use 18mm-equivalent only for city-establishing shots, 24mm for walking and one brief 2x portrait view for the food reaction. Natural handheld sway, small framing errors, autofocus adaptation and moderate depth so Tokyo remains readable. Warm amber stall light on faces, restrained cyan/magenta background accents, cool hair rim, deep detailed navy shadows and natural skin. 0–3s — RECOGNITION HOOK: 18mm elevated vertical view of the real multi-directional scramble crossing. City geometry fills at least 70% of frame; the host is a small figure. Include pedestrian signals, one compact taxi, station light boxes and wet reflections. Rapid human tilt toward street level. 3–6s — ENTER THE CITY: Hard cut to 24mm street-level handheld tracking as the same host and three friends cross with real pedestrian flow. She laughs once while walking. Keep the taxi line, diagonal markings and station facade coherent. 6–9s — YOKOCHO DISCOVERY: Hard cut beneath visible railway structure. A silver commuter train passes behind the group. Red lanterns, vending machines, noren, tiny counters and cables remain readable. She points once toward a grill stall while talking to the same friend. 9–12s — FOOD AND REACTION: Hard cut to food steam, then a brief 2x portrait as she tastes one bite and offers the remainder to a friend. Keep lanterns, vending-machine light and railway structure recognizable behind her. 12–15s — HERO ENDING: Hard cut behind the group as they exit toward the scramble crossing. She looks back over her left shoulder and gives one small “come with us” gesture. Preserve clean upper-left space for a post-production title. NATURAL MOVEMENT AND SOUND: Believable walking weight, hand-to-food contact, hair inertia, bag movement, steam and crowd occlusion. If native audio is supported, use only market chatter, grill sizzle, train pass, footsteps and natural laughter; final music and voiceover are added in post. CONTINUITY: Same host, friend group, wardrobe, bag, night, route, weather and color grade. Every transition is a hard cut; never dissolve or morph. FINAL VISUAL ACCEPTANCE CHECK: - Tokyo reads within the first second without text. - At least three shots retain three recognizable Tokyo anchors. - The same adult host, friends and wardrobe appear throughout. - Two facial views are flattering, realistic and free of wide-angle distortion. - Hands, food, walking and crowd movement remain physically plausible. - No readable generated text, identity drift or random location replacement. NEGATIVE PROMPT: generic Asian night market, Hong Kong streetscape, Bangkok market, Seoul shopping street, fantasy cyberpunk city, public-figure likeness, childlike face, identity drift, changing hair, changing clothes, duplicate friends, random new people, plastic skin, warped face, fisheye face, extra fingers, fused hands, impossible chopsticks, floating food, malformed train, train through buildings, empty street, overcrowded frame, influencer pose, professional gimbal, excessive shake, fake bokeh cutout, oversaturated neon, orange skin, crushed blacks, slow motion, dissolve, morph, captions, logo, watermark, readable focal text ``` ## First repair to try If the result becomes a generic neon market, change only the opening and environment-anchor language: increase the crossing, railway, taxi and station geometry coverage. Do not rewrite the character, wardrobe, action and lighting at the same time. --- FlareClip Prompt Lab · https://flareclip.com/prompt-lab Generated outputs remain subject to the terms of the model provider you use.