UlazAI - AI Image & Video Tools
Veo 3.1 Prompt Examples You Can Copy and Test
Every example below shows its prompt next to the real render it produced, with the model, video aspect ratio and camera cues listed beside it. Copy a prompt, watch the output, then send it straight into Veo 3.1 Studio.
Real prompts,
real renders,
copy and test.
Need the structure before the examples? Use the Veo 3 prompt guide for the official subject, action, scene, camera, style, audio and negative-prompt framework.
Need pricing or developer docs? View pricing or read the Veo 3.1 API docs.
Best Veo 3 prompt types for what?
Compare prompt examples by subject, camera, audio and use case, then adapt one in the prompt builder. Every prompt here is free to copy, whether you searched "veo 3 prompts" or typed it as one word ("Veo3 Prompts"); the cards list the action instructions and, where specified, the audio and camera cues. For the collection beyond these examples, browse the Veo prompts library.
Cinematic trailers
Use a shot type, one camera move, one lens cue, one light source, and one final emotional beat.
Browse cinematic examplesProduct ads
Name the product, material, scene setup, motion path, and exact reveal you want in the first seconds.
Launch the generatorSeries with one cast
Fix the cast, wardrobe and breed once, then change only location and action per clip so every scene stays on-model.
See the three-part seriesViral shorts
Use one core action, one surprise or payoff, short timing, and camera language that reads fast on mobile.
See pricing for production runs🔍 Search All Prompts
Searching...
Outputs are real renders from the UlazAI gallery and studio storage, published by their creators. The model that produced each render is named on the card; any prompt can be tested with Veo 3.1 Fast.
Product ads · 5 examples
UlazAI brand explainer
ULAZAI: highlight the core value proposition in an authentic, on-brand style. ULAZAI simplifies AI image and video generation by offering a single credit wallet for all premium models across its studios and API, eliminating tier lock-ins and per-model limits, and ensuring credits never expire. Create a clear demo/explainer story arc.
Use case: SaaS brand ad
Aspect: 16:9 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Hands the model the value proposition and asks for a demo arc, so the story sells one clear promise.
Test this promptTimes Square billboard reveal
A vertical 9:16 smartphone screen recording of a Facebook feed in French, static camera, no scrolling. The top sponsored post shows a huge photorealistic 3D anamorphic corner billboard in Times Square. On the billboard, a realistic red squirrel is standing on a tree branch surrounded by bright green leaves in sunny daylight. After a short pause, the squirrel suddenly jumps out of the billboard, breaking the screen plane in a convincing optical illusion, and lands on the white Facebook interface area below the post. It then runs toward the viewer across the phone screen, getting larger with natural perspective, realistic paws, detailed fur, and soft contact shadows. Small leaves and bark debris scatter across the interface as it moves. The city background and Facebook UI remain stable and readable while only the billboard content and squirrel animate. End with the squirrel very close to the camera, curious and slightly sniffing the lens, filling much of the lower center of the frame. Hyperrealistic viral social media ad style, bright daytime, smooth motion, no cuts, no camera pan, no extra animals, no surreal distortions, no cartoon look.
Use case: App or brand launch spot
Aspect: 9:16 · 8s
Camera: screen recording, static
Lens: anamorphic
Light: daylight
Audio: not specified
Why it works: Anchors a real location with neon and wet streets, then lets the billboard do the product reveal.
Test this promptFPV drone room makeover
From a micro FPV drone’s perspective, glide smoothly through a Nordic home in a cinematic way. Begin in an empty, dark room with no lights. As the drone sweeps around, quickly furnish the space with IKEA products in a fun and dynamic way, rug unrolls, sofa placed in a bouncy way, lamp and plants added like a symphony, all while the lighting gently rises from darkness to soft morning daylight, ending on a cozy, finished room and a symetrical end shot.
Use case: Furniture or interior brand ad
Aspect: 16:9 · 8s
Camera: aerial/drone, glide
Lens: not specified
Light: daylight
Audio: not specified
Why it works: A single drone path plus a timed transformation keeps the spot moving without cutting.
Test this promptTrade-fair photo becomes a spot
Create a short cinematic vertical video (9:16) from a real photo taken at the A+A trade fair. The image shows a man wearing a bright green jacket standing in front of a large green sign with the A+A logo and the printed text “19–22 October 2027 · Düsseldorf · Germany.” 🎯 Goal: Bring the existing photo to life — not redraw it. Use all visible text, the real A+A logo, and the event date exactly as they appear in the photo. Camera / Motion: • Begin with a slow cinematic zoom-in on the A+A logo and event date at the top of the sign. • Add gentle camera drift and parallax motion for depth and realism. • Then pan down smoothly to the man below, keeping the text legible throughout. Logo / Text Animation: • Apply a soft glow pulse to the real A+A logo (do not replace or redraw it). • Slightly animate the printed date “19–22 October 2027” with a subtle shimmer effect, keeping its original font and position. • Overlay clean animated text: 1️⃣ “SEE YOU AGAIN” – fade-in, bold white, soft outer glow. 2️⃣ “A+A 2027 IN DÜSSELDORF” – metallic shine, gentle float motion, spelled exactly as written. Visual Style: • Keep the original photo visible and realistic. • Add cinematic color grad
Use case: Event or B2B product ad
Aspect: 16:9 · 8s
Camera: zoom
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Turns one real event photo into motion while the reference keeps the product exactly as shot.
Test this promptLinkedIn B2B clip from a real photo
Use the uploaded image as the exact first frame and primary visual reference. Create an 8-second premium LinkedIn video about intelligent protective clothing and AI-supported textile services. Preserve the exact faces, identities, hairstyles, body proportions, clothing colors, garment details and group composition of the three people. Do not redesign the protective clothing and do not add any new logos. The video starts with the three people standing confidently in the original pose. Add only subtle natural movement: breathing, slight fabric movement, minimal posture adjustments and calm confident facial expressions. After one second, a soft red scanning line moves vertically across the three protective garments. As the scan passes, elegant transparent digital outlines briefly highlight pockets, stretch zones, reflective elements and protective sections of the clothing. The dark teal background slowly transforms into a subtle, futuristic textile-service environment. Show abstract conveyor movement, soft RFID signals, garment identification markers and intelligent sorting lights in the background. Keep all technical elements elegant, realistic and understated. Small particles
Use case: B2B social ad
Aspect: 9:16 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Starts the video on a real company photo, which keeps people, logo and office exactly on brand.
Test this promptUGC · 16 examples
Border collie picks up a mushroom
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch woman in her thirties with a dark ponytail, a mustard-yellow rain jacket, and her young black-and-white border collie. The woman kneels on the forest path holding her phone to her ear, worried but composed, the border collie sitting next to her, tail low.
Use case: Vet-warning content for pet brands
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: Pins the cast (owner, dog, clothing) once, then gives one sniff-and-grab action the camera simply follows.
Test this promptOwner calls after the mushroom grab
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch woman in her thirties with a dark ponytail, a mustard-yellow rain jacket, and her young black-and-white border collie. The border collie stops at the side of the path, sniffs and quickly snaps up a small mushroom from the leaf litter before the woman can react; she calls out and hurries closer.
Use case: Sequel clip in the same UGC series
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: Repeats the same CAST block so the dog and outfit stay identical while the scene moves forward.
Test this promptClose-up: photographing the mushroom
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch woman in her thirties with a dark ponytail, a mustard-yellow rain jacket, and her young black-and-white border collie. Close-up: the woman photographs the mushroom with her phone from above and then from the side showing the base, a leaf held next to it for scale.
Use case: Detail shot for an educational pet series
Aspect: 9:16 · 8s
Camera: close-up, handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: Keeps identity fixed and narrows to one macro action, which is how UGC series keep continuity cheap.
Test this promptDeath cap macro insert
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch woman in her thirties with a dark ponytail, a mustard-yellow rain jacket, and her young black-and-white border collie. Close-up of a pale greenish-white death cap mushroom (Amanita phalloides) with a bulbous base standing among oak leaves on the forest floor, morning light.
Use case: B-roll insert for a pet-safety story
Aspect: 9:16 · 8s
Camera: close-up, handheld
Lens: not specified
Light: morning light, natural light
Audio: ambient sound, music
Why it works: Names the exact species and the bulbous base, so the render has a verifiable subject instead of a generic mushroom.
Test this promptLump check under the jaw
Photorealistic handheld home video, vertical 9:16, natural light, shallow depth of field, real ambient room sound, no music, no speech, no text, no logos, no captions. Macro close-up under a golden retriever's jaw: fingertips part the fur and find a small firm rubbery lump the size of a marble, the dog stays relaxed, soft daylight.
Use case: Pet-health awareness ad
Aspect: 9:16 · 8s
Camera: close-up, handheld, macro
Lens: not specified
Light: daylight, natural light
Audio: ambient sound, music
Why it works: Describes fingertip action and a relaxed dog, so the emotion stays calm instead of drifting into drama.
Test this promptVet palpation in the clinic
Photorealistic handheld home video, vertical 9:16, natural light, shallow depth of field, real ambient room sound, no music, no speech, no text, no logos, no captions. Bright modern veterinary clinic: a female vet in a green scrub top gently palpates the neck of a golden retriever standing on the table while the owner strokes its back, reassuring mood, daylight.
Use case: Clinic or pet-insurance promo
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: daylight, natural light
Audio: ambient sound, music
Why it works: Adds the owner stroking the back to buy a reassuring mood inside an eight-second clip.
Test this promptLiving-room jaw check
Photorealistic handheld home video, vertical 9:16, natural light, shallow depth of field, real ambient room sound, no music, no speech, no text, no logos, no captions. A Dutch woman in her thirties kneels on a living-room rug beside a calm golden retriever and gently feels with her fingertips under the dog's jaw on both sides, the dog relaxed, soft morning window light, slow push-in.
Use case: UGC ad for at-home pet checks
Aspect: 9:16 · 8s
Camera: handheld, push-in
Lens: not specified
Light: natural light, window light
Audio: ambient sound, music
Why it works: Slow push-in plus morning window light gives one camera move and one light source, nothing to conflict.
Test this promptSofa check at dusk
Photorealistic handheld home video, vertical 9:16. A Dutch woman in her forties sits on a beige linen sofa in a warm living room at dusk, soft lamp light, and slowly strokes the flank of an older cream labrador lying beside her. Her fingers pause on a small soft lump under the fur; she looks down, calm and attentive. Natural skin and fur texture, shallow depth of field, gentle slow push-in, no text, no logos, no captions.
Use case: Senior-dog care content
Aspect: 9:16 · 8s
Camera: handheld, push-in
Lens: not specified
Light: dusk, lamp light
Audio: not specified
Why it works: The paused fingers and attentive look give the model a beat of acting rather than a list of objects.
Test this promptGum check, second angle
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch man in his fifties with short grey hair and a dark green wool sweater, and his old golden retriever with a light muzzle. The man at home on the sofa reads on a tablet with the golden retriever's head resting on his knee, warm evening lamp, calm and thoughtful.
Use case: Pet-health series insert
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: Same cast and rules with a NEW angle: how a series gets coverage without breaking identity.
Test this promptEvening phone note
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch woman in her thirties with a dark ponytail, a mustard-yellow rain jacket, and her young black-and-white border collie. Autumn forest path in the Netherlands, fallen leaves, the woman walks with her border collie off leash ahead of her, soft misty light, slow tracking.
Use case: Pet-health series, calm beat
Aspect: 9:16 · 8s
Camera: handheld, tracking
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: One prop action (typing a number) carries the story while the cast block keeps the dog identical.
Test this promptEmergency clinic at night
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch man in his fifties with short grey hair and a dark green wool sweater, and his old golden retriever with a light muzzle. The man sits at a wooden kitchen table with his golden retriever beside him, turning a small brown supplement jar in his hands, reading the label thoughtfully; morning light.
Use case: Pet-health series, night beat
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: morning light, natural light
Audio: ambient sound, music
Why it works: One walking action through a lit entrance; the cast block carries identity in low light.
Test this promptPowder on the food bowl
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch man in his fifties with short grey hair and a dark green wool sweater, and his old golden retriever with a light muzzle. Close-up of turkey tail mushrooms (Trametes versicolor) with brown, tan and cream concentric bands growing on a fallen oak log in a Dutch forest, dew on the moss, gentle camera drift.
Use case: Pet-supplement UGC ad
Aspect: 9:16 · 8s
Camera: close-up, handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: A measured spoon and a waiting dog turn a product step into a story beat.
Test this promptMushroom found on the path
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch man in his fifties with short grey hair and a dark green wool sweater, and his old golden retriever with a light muzzle. The man walks with his golden retriever along a quiet forest path in autumn light, the dog sniffing at a log with bracket fungi, the man watching calmly.
Use case: Pet-safety awareness clip
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: Names where the dog stops and what it does before the owner reacts, one cause, one effect.
Test this promptClinic check with the owner close
Photorealistic handheld home video, vertical 9:16, natural light, shallow depth of field, real ambient room sound, no music, no speech, no text, no logos, no captions. A woman sits on the sofa with her golden retriever's head on her lap, holding her phone to her ear making a calm call, notebook open beside her, afternoon light.
Use case: Vet-trust promo
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: Two calm actions in one frame (palpate, stroke) keep the mood reassuring, not clinical.
Test this promptTreat routine, same dog
Photorealistic handheld home video, vertical 9:16, natural light, shallow depth of field, real ambient room sound, no music, no speech, no text, no logos, no captions. Close-up from behind: a woman's hand feels behind the knee of a standing golden retriever's hind leg, the dog's tail wagging slowly, living room, natural light.
Use case: Pet-care series insert
Aspect: 9:16 · 8s
Camera: close-up, handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: The CAST block plus one tabletop action: another scene from the same on-model series.
Test this promptWater bowl check
Photorealistic handheld home video, vertical 9:16, natural light, shallow depth of field, real ambient room sound, no music, no speech, no text, no logos, no captions. Close-up: a woman's hands rest on the front of a golden retriever's shoulders where the neck meets the chest, fingers pressing lightly into the fur, the dog standing calmly on a rug, warm daylight.
Use case: Pet-daily-care content
Aspect: 9:16 · 8s
Camera: close-up, handheld
Lens: not specified
Light: daylight, natural light
Audio: ambient sound, music
Why it works: Names object, action and light so the ordinary moment still renders with intent.
Test this promptTikTok / Reels · 5 examples
Bedroom pop performance
Vertical 9:16 format video: Young woman in modern minimalist bedroom wearing pink athletic outfit standing centered singing upbeat pop song directly to camera, mouth moving expressively as if lip-syncing to fun playful lyrics, confident facial expressions, slight body sway and hand gestures, natural happy energy, soft natural lighting from windows, white walls and doors visible behind, authentic bedroom music content vibe, relatable singing session aesthetic AUDIO TYPE: Upbeat pop song (fun, playful, catchy chorus) TEXT TO ADD: "When the song hits different at 3am 🎵"
Use case: Music or fashion vertical
Aspect: 16:9 · 8s
Camera: not specified
Lens: not specified
Light: natural light
Audio: music, singing
Why it works: Centres one performer, one outfit and direct eye contact, the base grammar of a lip-sync vertical.
Test this promptSad folk on the couch
Vertical 9:16 video: Young woman on staircase singing dramatically to powerful ballad, hand on railing leaning forward with intensity, mouth clearly showing emotional delivery, passionate facial expressions, wearing casual outfit, dramatic side lighting creating shadows, stairwell architecture visible, theatrical music video aesthetic, intense emotional energy AUDIO TYPE: Powerful ballad (dramatic, emotional, builds to climax) TEXT TO ADD: "When you're home alone and channel your inner diva 🎭"
Use case: Storytelling or music vertical
Aspect: 16:9 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: music, singing
Why it works: Names the prop, the mood and the watery eyes, so the emotion is directable instead of random.
Test this promptStaircase ballad take
Vertical 9:16 video: Woman holding broom as microphone singing funky pop song, dancing around living room while cleaning, mouth forming fun lyrics, energetic movements and spins, casual outfit, bright living room, vacuum visible, pretending to clean but mostly performing, fun procrastination energy, comedy lifestyle moment AUDIO TYPE: Funky pop song (groovy, makes you want to dance, upbeat) TEXT TO ADD: "Cleaning the house takes 10 minutes, the concert takes 2 hours 🧹🎵"
Use case: Music vertical
Aspect: 16:9 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: singing
Why it works: A staircase gives vertical framing natural depth and one clear leaning pose to hold.
Test this promptBroom-microphone singalong
Vertical 9:16 video: Woman sitting on couch hugging pillow singing sad folk song, looking vulnerable at camera, mouth forming melancholic lyrics, slightly watery eyes, wearing oversized hoodie, cozy living room with dim lamp lighting, tissue box nearby, intimate emotional atmosphere, authentic heartbreak energy, relatable sad moment AUDIO TYPE: Sad folk/acoustic song (heartbreaking, vulnerable, slow) TEXT TO ADD: "It's been 3 years but this song still hurts 💔🎵"
Use case: Comedy music vertical
Aspect: 16:9 · 8s
Camera: not specified
Lens: not specified
Light: lamp light
Audio: singing
Why it works: One prop, one action, one eye-line: a joke the model can hold without a script.
Test this promptNear-miss suspense clip
Create a vertical realistic smartphone video built around suspense: the viewer should genuinely wonder whether the woman is about to be hit by the approaching car, but she must avoid it safely at the last second. A young athletic woman calmly holds a yoga tree pose barefoot on the double yellow lines of a steep palm-lined Los Angeles street. A white electric sedan approaches from behind her, getting visibly closer each second. She seems unaware, smiling and staying calm. Handheld viral phone footage, natural shake, real daylight, tense but funny. At the last possible safe moment, she suddenly loses balance and drops toward the sidewalk, landing safely on the curb as the car passes by without touching her. She laughs at the camera. No collision, no injury, no gore, just a near-miss illusion and viral suspense payoff.
Use case: Viral stunt-style vertical
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: daylight
Audio: not specified
Why it works: States the payoff rule first (no collision), then builds suspense, keeping the stunt safe and re-renderable.
Test this promptDialogue · 4 examples
Deadpan host with exact dialogue
Create a photorealistic vertical 9:16 social video, 7 to 8 seconds long. A dry, sarcastic elderly woman host, around 75 to 85 years old, sits in a cozy vintage living room filled with old furniture, lace curtains, porcelain decorations, warm lamp light, antique cabinets, and slightly cluttered old-fashioned details. She looks straight into the camera with sharp wit, subtle facial expressions, and calm confidence. Her tone is deadpan, slightly judgmental, clever, and memorable. She speaks fluent English with clear lip sync and natural pacing. Behind her, a television or monitor in the room shows a short moving preview or poster frame from the AI video she is introducing. The host briefly glances at that screen near the end. Camera is mostly locked medium shot, cinematic realism, natural skin texture, believable performance, no subtitles, no on-screen text, no watermark, no exaggerated gestures, no cartoon style. The video should feel surprising and scroll-stopping, like an unusual recurring character on social media. Her performance must feel witty and authentic, not overacted. Dialogue: "This little video cost less than a respectable lunch. Which is frankly insulting to the lunch. Look at it."
Use case: Recurring AI host character
Aspect: 9:16 · 8s
Camera: locked
Lens: not specified
Light: lamp light
Audio: dialogue
Why it works: Quotes the exact line and the lip-sync requirement, the only reliable way to get spoken words on target.
Test this promptTraffic-stop cameo with one line
Realistic body cam footage of a police officer pulling over a red Ferrari with mario driving. Mario only says "mama mia" and moves the head in fear, It was a serious offence, so the cop is extremely angry and tries to open the door of the car before Pikachu speeds away quickly. Miami, sunny day, palms
Use case: Meme or entertainment post
Aspect: 16:9 · 8s
Camera: body-cam
Lens: not specified
Light: not specified
Audio: dialogue
Why it works: Limits the character to a single quoted line, which keeps sync and comedic timing under control.
Test this promptStreet scene with spoken reactions
A staged, comedic street scene in New York City at late afternoon: a young rider on an absurdly small bicycle (wearing a helmet) casually rolls past two uniformed city police actors near a crosswalk, shouts a short censored remark [BEEP] from off-camera mic, the officers react in surprise and start jogging after him; handheld smartphone-style camera with a slight wide lens follows the rider weaving between pedestrians, cutting around a hot-dog cart and a mailbox, narrowly avoiding a taxi that brakes; the rider threads a gap between the two officers as they try to close in, then slips through and speeds off down the sidewalk, officers momentarily collide shoulder-to-shoulder and stop, breathing hard; crowd laughs; no injuries, no collisions, no brand logos, police insignia are generic City Police patches; mood playful, viral, witty; golden-hour light with long shadows, natural city ambience, distant siren and footstep foley; 9:16 vertical, ~12 seconds, 24 fps, sharp focus on the action, slightly desaturated street-doc color grade; clearly staged and safe.
Use case: Comedy skit
Aspect: 16:9 · 8s
Camera: handheld
Lens: not specified
Light: not specified
Audio: crowd
Why it works: Sets the stage and the passer-by reaction in one line each, so the joke lands in a single take.
Test this promptCCTV fail compilation
A series of raw, grainy CCTV clips stitched together like a fail compilation. Each feed has a timestamp flickering in the corner, wide fisheye distortion, and washed-out colors. In one scene, a man carrying a stack of boxes in a convenience store slips on a wet floor, scattering items across the aisle as startled customers react. Another clip shows a car misjudging a tight parking space, scraping a pillar in a dim garage under buzzing fluorescent lights. A third feed captures a person in an office stumbling over a chair, papers flying everywhere as coworkers laugh in the background. The footage has low frame rate, static interference, and compression artifacts, authentically mimicking real CCTV recordings. The audio is muffled and distorted, with echoes of startled shouts, clattering objects, and faint background hum. The palette is muted greys, pale greens, and cold fluorescents, emphasizing the raw, unpolished surveillance feel, shot in cinematic style.
Use case: Compilation-style content
Aspect: 16:9 · 8s
Camera: CCTV, fisheye
Lens: fisheye
Light: not specified
Audio: not specified
Why it works: Stitches multiple feeds with timestamps and fisheye distortion, a format viewers read as authentic instantly.
Test this promptCinematic · 9 examples
Chrome android in the rain
A chrome android steps out of a rain-soaked neon alley at night, water running down its shoulders, slow push in, cinematic
Use case: Sci-fi mood piece
Aspect: 16:9 · 8s
Camera: push-in
Lens: not specified
Light: neon, night
Audio: not specified
Why it works: One subject, one camera move and a wet-neon palette give the model a style it can actually hold for 8 seconds.
Test this promptDrone over misty peaks
Cinematic aerial drone shot soaring over misty mountain peaks at golden sunrise, volumetric god rays through clouds, nature documentary, ultra detailed
Use case: Nature intro or travel b-roll
Aspect: 16:9 · 8s
Camera: aerial/drone
Lens: not specified
Light: golden sunrise
Audio: not specified
Why it works: Names the light event (god rays at sunrise) rather than a vague mood, which is what sells aerial scale.
Test this promptAbstract tech background
Dark cinematic background with subtle yellow light particles floating. Abstract tech grid pattern slowly moving in the background. Dramatic, modern, premium feel. Dark navy blue tones with yellow accent glow. No text. 8 seconds.
Use case: Title-sequence background
Aspect: 9:16 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Explicitly bans text and logos, the difference between a usable background and a broken one.
Test this prompt10-second vertical fitness motivation video with hip hop
10-second vertical fitness motivation video with hip hop energy: SEC 0-2: Close-up of athletic woman's focused eyes and determined face, camera pulls back revealing her in starting position for workout in bright modern home gym, wearing black athletic outfit, gym equipment visible behind, she nods to camera with confidence, text slams in "LET'S GO 💯" in impact font, hype energy SEC 2-4: Quick cut to side angle, woman performing weighted squats with dumbbells, powerful controlled descent and explosive rise, muscles engaged, proper athletic form, mirror reflection visible in background, bright lighting, text appears "SQUAT IT OUT" with fire emoji, strength and power visible SEC 4-6: Fast transition with beat drop effect, woman now doing alternating jump lunges, explosive movements with air between reps, athletic power and agility, hair moving with motion, focused expression, text updates to "EXPLOSIVE POWER ⚡", high intensity workout energy SEC 6-8: Camera switches to front angle, woman in plank position transitioning to mountain climbers, rapid knee drives, core engaged, breathing visible sho
Use case: Cinematic b-roll
Aspect: 16:9 · 8s
Camera: close-up
Lens: not specified
Light: not specified
Audio: not specified
Why it works: A fully specified scene (light, movement, camera behaviour) that already rendered well.
Test this prompt<video version="2.0" model="sora-2" duration="8s" fps="2
<video version="2.0" model="sora-2" duration="8s" fps="24" aspect_ratio="9:16" resolution="1080x1920" seed="auto"> <look style="cinematic realism" color_grade="muted greens and grays, creamy whites, gentle rose skin" grain="subtle"/> <environment> <location>Back-alley stoop with green-painted steel stairs, cracked pavement, small strip of grass and stones</location> <time>Night</time> <weather>Dry pavement with slight humidity sheen</weather> <lighting> <key>On-camera flash pop look</key> <fill>Cool streetlamp spill from camera-left</fill> <background>Falls to shadow, retain highlight detail on white wings</background> </lighting> </environment> <camera sensor="smartphone-front feel" lens="26-28mm equiv" shutter="1/48s cinematic, mild motion blur" aperture="f/2.2 shallow DOF with soft halation on speculars" stabilization="handheld"> <movement type="single push-in" path="straight ~0.5m" micro_tilt="minimal to keep face centered" wobble_limit="handheld micro-bob ≤ 3% frame height"/> <rules>One primary motion only, no zooms, no extra moves, no cuts</rules> </camera> <cast> <human id="subject0
Use case: Cinematic b-roll
Aspect: 16:9 · 8s
Camera: handheld, push-in, tilt
Lens: not specified
Light: night
Audio: not specified
Why it works: A fully specified scene (light, movement, camera behaviour) that already rendered well.
Test this promptCinematic 16:9 format, 10-second premium promo video: SE
Cinematic 16:9 format, 10-second premium promo video: SECONDS 0-3 (HOOK): Slow-motion aerial drone shot rising over stunning mountain landscape at golden hour, warm orange and pink sky, camera rises up revealing valley below, epic cinematic scale, text fades in upper center "THE FUTURE OF VIDEO CREATION" in elegant white serif font, lens flares and god rays, inspirational atmosphere, professional film quality SECONDS 3-6 (TRANSITION): Image ripples and transforms into digital portal effect, real landscape transitions to floating holographic UlazAI platform interface hovering in elegant dark void, platform made of glass and chrome materials with reflections, multiple video thumbnails playing on the interface showing Sora 2 content, camera slowly orbits around platform, previous text fades out, new text appears "SORA 2 • AVAILABLE NOW" in modern sans-serif font center screen, premium tech aesthetic SECONDS 6-10 (CLOSE): Camera continues orbit, platform rotates to face camera directly, zooms in slightly to featured video playing on interface, text updates to "ULAZAI.COM/SORA" in large bold
Use case: Cinematic b-roll
Aspect: 16:9 · 8s
Camera: aerial/drone, orbit, zoom
Lens: not specified
Light: golden hour
Audio: not specified
Why it works: A fully specified scene (light, movement, camera behaviour) that already rendered well.
Test this prompt{ "model": "Veo3Fast", "duration_seconds": 9, "aspect_ra
{ "model": "Veo3Fast", "duration_seconds": 9, "aspect_ratio": "16:9", "prompt": "Interior of a realistic European city bus in daytime. Blue patterned seats and yellow vertical poles. Two human passengers sit mid-front, facing camera, composed and slightly awkward. Several small macaque monkeys are seated like commuters and hold onto poles, with subtle idle motions such as blinking and gentle head turns. Tone is dry comedic but respectful. Lighting is soft daylight from side windows. Start with a rear-aisle view and slowly dolly forward to a medium two-shot of the two humans, keeping the monkeys visible in surrounding seats. Keep movements natural and small. No text on screen.", "camera": { "movement": "slow dolly in from rear aisle to mid-bus two-shot", "focal_length": "35mm", "aperture": "f/2.8", "frame_rate": 24, "motion_strength": "low" }, "style": { "look": "cinematic realism", "color": "neutral with gentle contrast", "grain": "light" }, "audio_notes": { "sfx": [ {"timecode": "00:07.5", "cue": "short bus PA chime"} ], "voiceover": { "language": "nl-NL", "voice": "male driver,
Use case: Cinematic b-roll
Aspect: 16:9 · 8s
Camera: dolly
Lens: 35mm
Light: daylight
Audio: voice
Why it works: A fully specified scene (light, movement, camera behaviour) that already rendered well.
Test this promptCreate a 10–12 second vertical video set in a modern gym
Create a 10–12 second vertical video set in a modern gym. Extremely fast cuts. Start calm: a friendly female client (beginner) walks with a personal trainer and asks, ‘Where’s the machine I used yesterday?’ The trainer nods seriously and points to the vending machine filled with snacks and drinks. Whip-pan reveal, tiny beat drop; she smiles and starts toward it. Then the trainer laughs, points to the real machine (leg press or rower), and they walk there together. Close-ups: smiles, high-five, first rep. Music: upbeat, cinematic. Camera: whip pans, rack focus, macro inserts, short time-ramps. Pacing: question (calm) → vending reveal (twist) → positive payoff. No on-screen text or captions.
Use case: Cinematic b-roll
Aspect: 16:9 · 8s
Camera: close-up, macro
Lens: not specified
Light: not specified
Audio: music
Why it works: A fully specified scene (light, movement, camera behaviour) that already rendered well.
Test this prompt<video model="sora-2" duration="10s" fps="24" aspect_rat
<video model="sora-2" duration="10s" fps="24" aspect_ratio="16:9"> <prompt> Modern gym, treadmills and mirrors, soft light. In the far corner stands an unbranded snack vending machine. A plus size woman asks a personal trainer: "Where is the machine I used yesterday?" The trainer points toward the vending machine. Camera makes a slow dolly from a two-shot to the corner reveal. She smiles and walks to the vending machine. Hold a brief beat as she reaches for the panel, then fade out. Realistic, light comedic misunderstanding. No text on screen or logos. </prompt> </video>
Use case: Cinematic b-roll
Aspect: 16:9 · 8s
Camera: dolly
Lens: not specified
Light: soft light
Audio: not specified
Why it works: A fully specified scene (light, movement, camera behaviour) that already rendered well.
Test this promptImage-to-video · 2 examples
Slider moves, horse runs
Make the slider move and the horse running and moving. Keep the rest static.
Use case: Animate a still product or scene
Aspect: Auto · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Says what must move and what must stay static, the core discipline of image-to-video prompts.
Test this promptStand up and drill
Make him stand up and start drilling in the wall behind him
Use case: Animate a person from a photo
Aspect: 16:9 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: One subject, one new action, one location detail: enough direction without fighting the source image.
Test this promptCharacter consistency · 3 examples
One CAST block, multiple scenes
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): a Dutch man in his fifties with short grey hair and a dark green wool sweater, and his old golden retriever with a light muzzle. Close-up of the man's hands sprinkling a small measured spoon of brown mushroom powder over a bowl of dog food, the golden retriever waiting patiently.
Use case: Series production
Aspect: 9:16 · 8s
Camera: close-up, handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: Declares wardrobe, age and breed once and reuses it in every shot, which is how the whole series stays on-model.
Test this promptTeam photo as first frame
Use the uploaded team photo as the exact visual reference and first frame. Create an 8-second premium B2B LinkedIn video about innovation in protective workwear. Preserve the exact identity, facial features, body proportions, hairstyles, workwear design, navy-blue colors, reflective details and all visible garment elements of the three people. Do not replace or redesign the clothing. The video begins with the three people standing confidently in the same formation as in the reference image. Hold the original composition for the first half second. Then create a very slow, smooth cinematic dolly-in. Add only subtle, realistic human movement: natural breathing, slight posture adjustments, minimal fabric movement and small confident changes in facial expression. Nobody speaks and there is no lip movement. A thin, elegant red scanning light moves slowly from left to right across the protective clothing. As the scanning light passes, tasteful semi-transparent technical interface elements briefly highlight stretch zones, reflective piping, pockets and protective garment construction. The interface graphics must remain abstract and premium, with no readable text and no fake logos. In
Use case: B2B brand video
Aspect: 9:16 · 8s
Camera: dolly
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Uses the uploaded photo as exact visual reference, so real people stay recognisable instead of being reinvented.
Test this promptProduct as fixed reference
Create a premium product-focused video using the provided glove image as the main visual reference. Keep the glove design exactly consistent: red fine-knit cut-resistant glove with black sandy nitrile foam coating, long red cuff with blue trim, product print clearly visible, realistic proportions, no distortion. Animate the glove as the hero object in the foreground with subtle parallax and a slow cinematic push-in. The glove should remain sharp, centered and dominant, while the background becomes a realistic industrial motion scene inspired by the glove’s application and features. Background theme: modern metalworking and industrial manufacturing environment, with soft blurred elements such as sheet metal, steel components, sharp-edged parts, glass panels, workshop surfaces and industrial structures. Add subtle motion in the background: slow moving metallic reflections, light sweeps, drifting particles, soft industrial haze, and abstract motion suggesting cutting-risk environments without showing danger or injury. Visually communicate the glove’s core strengths: - extreme cut protection - secure grip - industrial durability - handling of sharp-edged materials - premium professi
Use case: Product ad with a real product
Aspect: 16:9 · 8s
Camera: push-in
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Tells the model to keep the glove design exactly consistent, the key rule for product-True renders.
Test this promptMulti-shot · 3 examples
Part 1: the gum check in the park
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): an East Asian woman around sixty with short grey hair, a light blue rain jacket and black trousers, and her large adult german shepherd. Close-up in a park: the woman kneels in the grass and gently lifts the upper lip of her german shepherd lying on its side to look at its gums; the gums look pale; her face shows worry; soft afternoon light.
Use case: Three-part story, clip 1
Aspect: 9:16 · 8s
Camera: close-up, handheld
Lens: not specified
Light: natural light
Audio: ambient sound, music
Why it works: The CAST block plus one location and one action makes this the cleanest possible first shot of a series.
Test this promptPart 2: saving the vet number
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): an East Asian woman around sixty with short grey hair, a light blue rain jacket and black trousers, and her large adult german shepherd. Evening at a kitchen table: the woman types an emergency vet clinic phone number into her phone and saves the contact, the german shepherd lying at her feet, warm lamp light.
Use case: Three-part story, clip 2
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: lamp light, natural light
Audio: ambient sound, music
Why it works: Same cast, new location, one prop action: a continuation formula that holds identity across clips.
Test this promptPart 3: the clinic entrance
Photorealistic handheld video, vertical 9:16, natural light, shallow depth of field, real ambient sound, no music, no speech, no text, no logos, no captions. CAST (same in every shot): an East Asian woman around sixty with short grey hair, a light blue rain jacket and black trousers, and her large adult german shepherd. Evening: the woman walks her german shepherd quickly through the lit glass entrance of a veterinary emergency clinic, calm and determined, warm light from inside.
Use case: Three-part story, clip 3
Aspect: 9:16 · 8s
Camera: handheld
Lens: not specified
Light: natural light, warm light
Audio: ambient sound, music
Why it works: Closes the arc with a single walk-in action and warm interior light for a natural ending frame.
Test this promptBad prompt to better prompt · 3 pairs
From "bird flying" to a directed loop
Bad prompt · Veo 3.1 Lite
bird flying
No action, no light, no framing: the model has to guess everything, so results are random.
Better prompt · Veo 3.1 Fast
[4K] make the bird fly away in the sunset
Names the action (fly away), the light (sunset) and a quality bar, so the render has a target.
Use case: Prompt writing lesson
Aspect: 9:16 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Names the action (fly away), the light (sunset) and a quality bar, so the render has a target.
Test this promptFrom "make it move" to a sequenced i2v instruction
Bad prompt · Veo 3.1 Fast
Make the slider move and the horse running and moving. Keep the rest static.
"Move it" without naming what stays static invites warping and drifting props.
Better prompt · Veo 3.1 Fast
Premium realistic tutorial video. Adult hands use a blue diamond painting pen at a bright wooden craft table. First the pen tip touches a small pink wax pad, then the pen picks up one blue rhinestone diamond from a white grooved tray and moves toward the adhesive canvas. Warm daylight, clean Italian ecommerce style, gentle camera movement, no on-screen text.
Orders the hands (wax, pick, place) and locks the scene, which keeps a tutorial usable.
Use case: Prompt writing lesson
Aspect: 16:9 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: Orders the hands (wax, pick, place) and locks the scene, which keeps a tutorial usable.
Test this promptFrom one crowded line to one subject, one emotion
Bad prompt · Veo 3.1 Fast
kat en hond in een auto race met een muis
A cat, a dog, a car race and a mouse in eight seconds is a collision of subjects the model must average.
Better prompt · Veo 3.1 Fast
A stressed contractor at a messy desk covered in papers and sticky notes. He rubs his temples in frustration while his phone rings. Warm office lighting, realistic cinematic style.
A single character with a named emotion and warm office light renders clean and casts well.
Use case: Prompt writing lesson
Aspect: 9:16 · 8s
Camera: not specified
Lens: not specified
Light: not specified
Audio: not specified
Why it works: A single character with a named emotion and warm office light renders clean and casts well.
Test this promptHow to Use These Examples
✅ Best Practices
- • Copy the exact structure but modify details for your needs
- • Combine elements from different examples
- • Start with simpler examples and build complexity
- • Pay attention to the technical terms used
- • Test variations to see what works best
⚠️ Common Mistakes
- • Don't copy prompts without understanding the elements
- • Avoid mixing conflicting styles or moods
- • Don't make prompts too long or complex
- • Don't ignore the difficulty level recommendations
- • Don't expect perfect results on first try
🚀 Ready to test these Veo 3 prompts in production?
Take the prompt pattern you want, then move straight into Veo 3.1 Studio, the AI Prompt Assistant, pricing, or the API reference. This page should help you go from example to output without leaving the UlazAI flow.
Veo 3.1 JSON prompt structure
For multi-shot and repeatable results, Veo 3.1 responds well to a structured JSON prompt instead of one long sentence. The plain-language template above still works; JSON just makes style and continuity explicit so each scene stays consistent.
{
"output": { "aspect_ratio": "16:9", "duration_seconds": 8 },
"global_style": "cinematic, shallow depth of field, warm key light",
"continuity": "same character, same wardrobe across all scenes",
"scenes": [
{
"subject": "a barista in a small specialty coffee bar",
"action": "pulls an espresso shot, steam rising",
"camera": "slow push-in on the portafilter",
"audio": "milk steaming, low cafe murmur"
},
{
"subject": "the same barista",
"action": "slides the finished cup across the counter",
"camera": "handheld over-the-shoulder",
"audio": "cup taps the saucer, soft jazz"
}
],
"negative_prompt": "text overlays, distorted hands, logo"
}
Map each key to the same idea as the sentence formula: subject + action + camera + style + audio, plus a negative_prompt to remove artifacts. Paste it into the Veo 3 prompt assistant or send it to the Veo 3.1 API.
Veo 3 Prompt FAQ
Are the renders on this page made with Veo 3.1?
Most are, and every card names the model that produced its render. Prompts that originally ran on another model test just as well in Veo 3.1 Fast.
Why do the UGC examples repeat the same cast block?
The CAST line fixes age, clothing and breed once, then each clip only changes location and action. That is what keeps a series on-model.
Can Veo 3 prompts be reused across campaigns?
Yes. Build a prompt base template for style and camera and add one change per iteration to maintain consistency across variants.
How do I improve Veo 3 results?
Improve by splitting long prompts, naming a consistent atmosphere, and adding exact technical direction like movement rhythm and shot duration.
Building this into your own product? See the Sora 2 API documentation and the Kling 3.0 API documentation.
Copy a Veo prompt only after checking its shot logic
A good example keeps subject, action, camera and sound compatible within one short scene. Replace the subject and brand details, but preserve the timing and camera relationship that made the example work.
Use JSON examples as editable structure, not proof of an official Veo format. Check the chosen model’s current inputs before generating.