Streetwear lookbooks fall apart the moment the model starts walking like a runway model. Streetwear is not couture: the pose, gait, and energy have to read “hanging out,” not “presenting a garment.” The default failure mode is a stiff catwalk strut, an empty backdrop, and a hard rim light that screams editorial. The 10 templates below lock a relaxed pace, real urban props, and a per-look palette so each beat in the lookbook still looks like the same brand world.
All 10 prompts are written for the current generation of text-to-video and image-to-video models (June 2026): Google Veo 3.1, Kling 3.0, and Runway Gen-4.5. The prompt language is model-agnostic, but pick the right tool for the job before you start — see the comparison table below.
TL;DR
- A usable streetwear clip needs all five layers: lens, light state, named slow camera move, a two-color per-look palette, and one motion per 5-8 second clip.
- For a 10-look set, lock the same lens and clip length across every beat and let only the palette change. That is what makes ten clips read as one campaign.
- Use an image-to-video model (Kling 3.0 or Runway Gen-4.5) with a single reference still to hold the model’s identity across looks. Text-only prompts drift.
- Base clips top out near 8 seconds on Veo 3.1; Kling 3.0 reaches ~15 seconds with multi-shot. Plan beats accordingly.
Which model to use
Every model on this list runs from the same prompt vocabulary, but they trade off cost, clip length, and how well they hold a face across looks. As of June 2026:
| Model | Best for | Base clip | Native res | Cost (API) | Identity hold |
|---|---|---|---|---|---|
| Kling 3.0 | Value + image-to-video; strongest at reducing body/face warp | ~3-15s (multi-shot) | up to 4K | ~$0.10/sec | Strong (3D face/body reconstruction) |
| Runway Gen-4.5 | Precise camera control + reference-driven consistency | ~5-10s | up to 4K | ~$0.15/sec | Strong (reference image support) |
| Veo 3.1 | Safe all-rounder, native audio, keyframe start/end | 8s (extends in chunks) | 720p/1080p/4K | ~$0.15/sec fast | Good (multi-image reference) |
| Sora 2 | Story-led, multi-shot narrative | variable | up to 4K | ~$0.75/sec | Good |
Practical picks: for a budget lookbook where holding the same model matters most, start with Kling 3.0 in image-to-video. For granular camera moves (the slow dollies and tracking shots these templates lean on), Runway Gen-4.5. Veo 3.1 is free to try via Google Labs and Gemini, and is included with Google AI Pro at $19.99/month. Pricing and clip limits move fast in this category; confirm current numbers on the Veo developer pricing page and the Runway pricing page before you commit a budget.
What a high-quality video prompt should contain
Five layers, every time:
- Lens:
35mm,50mm, occasional24mm widefor street context - Light state:
late afternoon,neon sign rim,overcast soft,tungsten storefront spill - Camera motion: named slow:
slow dolly in,slow tracking from behind,static medium - Color palette: per look:
concrete grey + safety orange,neon magenta + cyan,washed denim + cream - Subject restraint: one move per clip, 5-8 second clip
If the prompt skips any of the five, the model fills the gap with random motion and the lookbook stops looking like one campaign. Keep each clip at or under the model’s base length (8s on Veo 3.1, ~15s with Kling 3.0 multi-shot) so you are not relying on a shaky scene-extension chain.
10 copy-ready video prompt templates
1. Graffiti wall slow strut
Best for: drop teaser for a graphic tee or hoodie
A young man in an oversized graphic hoodie walks slowly past a graffiti-covered concrete wall, hands in pockets, head slightly down, relaxed gait. 35mm lens, slow lateral tracking from the side, late afternoon soft light, concrete grey and saturated graffiti color palette, 6-second clip, no camera shake.2. Subway platform, train passing
Best for: techwear or transit-themed capsule
A model in techwear stands still on a near-empty subway platform, looking just past camera. A train passes behind with motion blur and warm interior light streaming. 35mm lens, static medium shot, single fluorescent overhead, teal and warm amber palette, 7-second clip.3. Skate park trick stand
Best for: skate-adjacent streetwear, mid-tier price point
A model in baggy carpenter jeans and a vintage tee stands beside a skateboard on the lip of a concrete bowl, one foot on the deck, looking off-frame. 35mm lens, slow handheld push in, overcast diffused daylight, washed denim and faded concrete palette, 6-second clip.4. Rooftop wind hair flow
Best for: aspirational lookbook hero clip
A model on an empty city rooftop at golden hour, oversized open shirt and tank, wind moves the shirt and hair softly, eyes off-frame. 50mm lens, very slow dolly in from waist up, warm orange and cool teal sky palette, 7-second clip, no other people in frame.5. Neon convenience store doorway
Best for: late-night drop, capsule with reflective trims
A model leans casually in the open doorway of a 24-hour convenience store at night, holding a paper cup. Neon magenta and cyan storefront signs spill onto the sidewalk. 35mm lens, static medium shot with very slow zoom in, magenta and cyan palette, 6-second clip.6. Basketball court late afternoon
Best for: court-culture capsule, retro athletic
A model in retro athletic shorts and a vintage jersey stands center court at an empty outdoor basketball court, holding a ball at the hip, looking past camera. 35mm lens, slow tracking right behind chain-link fence in foreground, late afternoon warm side light, faded orange and cracked-paint blue palette, 7-second clip.7. Vintage thrift shop entry
Best for: secondhand or upcycled brand storytelling
A model pushes open the glass door of a small vintage thrift shop and steps in, head turning slightly to look at the racks. 35mm lens, static medium from inside the shop looking out, warm tungsten interior plus cool daylight from outside, washed cream and dusty pink palette, 6-second clip.8. Downtown crosswalk back-shot
Best for: graphic-back hoodie or jacket reveal
A model walks slowly away from camera across a downtown crosswalk during a green light, crowd blurred in foreground, oversized hoodie graphic on the back fully readable. 50mm lens, static low shot from curb height, overcast diffused daylight, muted grey and asphalt-black palette, 6-second clip, single subject in focus.9. Parking garage neon dolly
Best for: techwear or futurist drop teaser
A model in a long techwear coat stands alone in a dim concrete parking garage, single neon strip light glowing magenta along the ceiling behind. 35mm lens, slow dolly in straight toward subject, deep contrast, magenta rim and deep teal shadow palette, 6-second clip, no movement other than slight head tilt.10. Alley brick wall lean
Best for: classic streetwear portrait beat in a multi-look lookbook
A model leans casually against an old red brick alley wall, one knee bent and foot against the wall, hands in jacket pockets, looking just off-camera. 50mm lens, static medium shot, soft late afternoon bounced light, brick red and warm cream palette, 5-second clip.Common mistakes
- Letting the model walk like a runway model. Streetwear gait is loose, not pulled-up.
- Empty backdrops everywhere. Streetwear needs urban context: graffiti, signage, fencing, concrete.
- One generic palette across all looks. Each beat should have its own grade.
- Logos in the prompt. They almost always warp, even on Kling 3.0 and Veo 3.1. Leave room for post compositing instead.
- Multi-action clips (“walks then turns then jumps”). Pick one motion per 5-8 second clip; chained actions are where physics breaks first.
- Text-only identity. If the same model has to appear across 10 looks, feed the model a reference still in an image-to-video flow rather than re-describing the face every time.
How to push results further
- Per-look palette grading is the single biggest upgrade: each clip gets its own two-color recipe.
- Replace fluorescent overheads with practical light (storefront spill, neon strips, a single sodium street lamp) for instant realism.
- Add
relaxed gait, hands in pockets, head slightly downto break the runway-stride default. - Cast urban props as supporting subjects (skateboard, chain-link fence, basketball, paper cup) instead of as scenery.
- Lock one lens plus one clip length across the lookbook (for example, always 35mm and 6 seconds) so the cut feels like one campaign.
- Use Veo 3.1 keyframe control to set a start and end frame on hero beats; it stabilizes the move far better than a pure text prompt.
- For matching audio (ambient street noise, distant traffic), Veo 3.1 and Kling 3.0 both generate native sound, so you can skip a separate sound pass on social cuts.
FAQ
Q: Why does my streetwear model still walk like a couture model?
A: Add explicit gait language — relaxed gait, loose walk, hands in pockets, head slightly down. Without these, video models default to a runway stride.
Q: How long should each lookbook beat be?
A: 5-7 seconds. Lookbook beats are short and many; a cinematic single-shot 10s+ take is for ad films, not lookbooks. This also keeps you inside the 8-second base length on Veo 3.1, so you avoid scene-extension artifacts.
Q: Which model holds the same model’s face best across 10 looks?
A: As of June 2026, image-to-video on Kling 3.0 (3D face/body reconstruction) and Runway Gen-4.5 (reference image support) hold identity best. Generate one clean reference still, feed it as the reference frame, then add a tight verbal description per beat. Pure text-to-video drifts noticeably by the third or fourth look.
Q: What does it cost to make a full 10-look set?
A: At API rates (June 2026), a 6-second clip is roughly $0.60 on Kling 3.0 ($0.10/sec) or $0.90 on Runway Gen-4.5 ($0.15/sec), before retries. Budget for 2-3 takes per look, so a 10-look set lands around $15-30 of generation plus your editing time. Veo 3.1 is free to try via Google Labs and Gemini before you commit to paid API runs.
Q: How do I make the lookbook feel like one brand?
A: Lock lens, clip length, and gait vocabulary across all 10 beats. Let the palette vary per look, but keep the structure constant.
Q: Why do my night neon clips look muddy?
A: Specify only two neon colors (magenta and cyan, never three or more) and add deep contrast, no ambient fill. Models smear when given too many neon hues to balance.
Related articles
- Fashion Runway Video Prompts: 10 Editorial-Grade Templates
- Cinematic AI Video Prompt Examples
- Editorial Fashion Video Prompts: 10 Magazine-Style Templates
- Social Ad Hook Video Prompts: 10 “3-Second Stop-the-Scroll” Templates
- All AI Video Prompts
Tags: #Fashion #streetwear #lookbook #Video generation #Prompt