AI Video Prompts: 42 Copy-Paste Examples + the Formula That Works (2026)

The best AI video prompts describe a shot, not a picture: who or what is on screen, the one action that happens, how the camera moves, where it takes place, and the light and style. Below is the structure that works across Veo, Seedance, Wan and most other video models, plus 42 copy-paste AI video prompts sorted by category, including image-to-video prompts. Updated as of 18 October 2026.

What makes a good AI video prompt?

A video model has to invent motion over time, so a prompt that only describes a still frame leaves it guessing. Google's Veo 3.1 prompting guide breaks a prompt into five parts: cinematography, subject, action, context, and style and ambiance. Adobe's Firefly guide suggests almost the same order: visual style, shot type, character, action, location, aesthetic. They name the parts differently, but they cover the same ground.

Two habits matter more than adding adjectives:

1. One clear action per clip. Stacking three events into a 5-second clip usually blurs all of them. Several guides also suggest pairing the main action with one or two small background motions (steam, hair in wind, passing traffic).
2. Describe what you want, not what you don't. Google's guide recommends "a desolate landscape with no buildings or roads" style phrasing over a bare "no man-made structures" negative.

How long should an AI video prompt be?

For a 4–10 second clip, 40–80 words is enough to cover the six parts above without contradicting yourself. HeyGen's guide notes that one sentence is fine for quick drafts, while a reusable ad or explainer may need a full brief with a script and style notes. If a result is wrong in one way, edit that one part of the prompt instead of rewriting the whole thing.

Which camera words do AI video models understand?

These terms show up in official guides (Google, Adobe) and are reliably understood by current models:

• Framing: extreme close-up, close-up, medium shot, wide shot, aerial / drone shot, over-the-shoulder, POV, two-shot
• Movement: static / locked-off, slow pan, tilt up, dolly-in, dolly-out, tracking shot, crane up, orbit, handheld
• Lens & focus: macro lens, shallow depth of field, rack focus, deep focus, anamorphic
• Pacing: slow motion, real-time, time-lapse

Pick one movement per shot. "Slow dolly-in while orbiting and tilting" tends to produce wobbly camera paths.

42 copy-paste AI video prompts by category

Each prompt follows the shot → subject → action → setting → style order. Swap the nouns and keep the structure. All are written for 5–8 second clips.

Cinematic prompts

1. Wide aerial shot, slow push forward over a misty pine forest at sunrise, a single red train crossing a stone viaduct, golden light breaking through fog, epic film look, anamorphic lens flare.
2. Close-up, slow dolly-in on an old lighthouse keeper's face as he looks out to sea, wind moving his grey beard, overcast stormy light, muted teal and orange grade, 35mm film grain.
3. Low-angle tracking shot following a woman in a long red coat walking through a crowded rainy Tokyo crossing at night, umbrellas around her, neon reflections on wet asphalt, shallow depth of field.
4. Static wide shot of a lone astronaut standing on a red desert dune, dust blowing across the frame, twin suns setting on the horizon, science-fiction epic, soft haze.
5. Handheld medium shot, a chef flambés a pan in a busy restaurant kitchen, flames flare up and light his face, steam and movement in the background, warm tungsten light, documentary style.
6. Crane up from a quiet village square to reveal snow-covered mountains behind it, church bell swinging, smoke rising from chimneys, blue-hour light, calm and wide.
7. Slow-motion close-up of a dancer's hand releasing a cloud of coloured powder against a black background, particles drifting, rim light, high-speed camera look.

Product and ad prompts

8. Macro shot, slow orbit around a glass perfume bottle on a black mirror surface, a single drop of water rolls down the glass, soft studio rim light, luxury commercial look.
9. Top-down static shot, a pair of white sneakers drops onto a pastel pink floor and bounces once, small dust puff, clean studio lighting, playful commercial style.
10. Medium shot, slow dolly-in on a ceramic coffee mug on a wooden café table, steam curling up, morning sunlight through a window, cosy lifestyle ad, shallow depth of field.
11. Close-up tracking shot along a wireless headphone set rotating on a turntable, light sweeps across the matte surface, dark gradient background, tech product reveal.
12. Wide shot, a skincare bottle sits on wet river stones as clear water flows around it, leaves drifting past, soft natural daylight, fresh and minimal.
13. Static shot of a smartphone on a marble desk, the screen lights up and a notification slides in, hands enter frame to pick it up, bright clean office light.
14. Slow-motion close-up, a splash of orange juice bursts around a falling orange slice, droplets frozen in the air, white background, crisp commercial lighting.

Travel and landscape prompts

15. Drone shot gliding low over turquoise water toward a white-sand beach, small waves breaking, palm trees swaying, bright midday sun, travel-film look.
16. Time-lapse wide shot of clouds rolling over Icelandic mountains, the shadow line racing across green valleys, moody overcast light.
17. Tracking shot from inside a train window, terraced rice fields passing by, a farmer in a straw hat looks up, soft afternoon light, gentle motion blur.
18. Slow pan across a Moroccan spice market, colourful cones of spices, vendors chatting, sunbeams through a fabric canopy, warm documentary style.
19. Static wide shot of the northern lights rippling over a frozen lake, a small cabin with glowing windows, stars visible, long-exposure look.
20. POV walking shot through a narrow Santorini alley, white walls and blue doors, a cat crosses the path, bright Mediterranean sunlight.
21. Aerial orbit around a desert canyon at golden hour, a small hiker on the ridge, long shadows, warm red rock tones.

Food prompts

22. Macro close-up, honey pours slowly onto a stack of pancakes, butter melting and sliding down, warm morning window light, food-commercial style.
23. Top-down shot, hands sprinkle grated parmesan over a bowl of steaming pasta, steam rising, rustic wooden table, natural light.
24. Slow-motion close-up of a burger being pressed together, sauce dripping, sesame seeds bouncing, dark moody background, rim light.
25. Medium shot, a barista pours latte art into a cup, a heart shape forms, café noise blur in the background, soft warm light.
26. Close-up dolly-in on a ramen bowl as chopsticks lift noodles, steam swirling, soft-boiled egg glistening, cinematic shallow depth of field.
27. Static macro shot of chocolate melting over a strawberry, a slow drip falls from the tip, glossy texture, black background.
28. Wide shot of a street-food grill at night, flames flaring under skewers, smoke drifting through string lights, lively handheld feel.

Anime and stylised prompts

29. Anime style, wide shot, a swordswoman with short silver hair stands on a rooftop at sunset, her long travelling coat flutters in the wind, cherry petals drift past, soft cel shading.
30. Studio-Ghibli-inspired watercolour style, a small fox runs through tall grass toward a hilltop windmill, clouds drifting, gentle breeze in the grass.
31. Cyberpunk anime, low-angle shot of a hover bike speeding through a neon city canyon, rain streaks, holographic signs flicker, dynamic motion lines.
32. 3D animated style, a round robot waters a tiny plant on a windowsill, the leaf perks up, the robot's eyes blink happily, soft pastel lighting.
33. Claymation style, close-up of a clay snail sliding across a mossy log, leaving a shiny trail, stop-motion texture, warm forest light.
34. Paper cut-out animation, a whale swims through layered paper waves, small boats bobbing above, slow parallax camera move.
35. Pixel-art style, side-scrolling shot of a knight walking through a forest at night, fireflies blinking, gentle parallax background.

Vertical social video prompts (9:16)

36. Vertical 9:16, handheld selfie-style medium shot, a woman in a cosy sweater holds up a book and smiles at the camera in a bright bedroom, natural window light, casual creator style.
37. Vertical 9:16, top-down shot of hands unboxing a small gift box, tissue paper lifts to reveal a silver necklace, soft daylight on a white bedsheet.
38. Vertical 9:16, POV shot pouring iced coffee into a glass full of ice, milk swirling, kitchen counter, bright morning light, satisfying slow pace.
39. Vertical 9:16, static shot of a man jogging toward the camera along a seaside path at sunrise, breath visible in cool air, energetic fitness style.
40. Vertical 9:16, close-up of a cat batting at a dangling toy, quick paw swipes, cosy living room, warm lamp light, funny pet clip.

How do you write image-to-video prompts?

When you start from an image, the picture already fixes the subject, colours, lighting and composition. Runway's help center says the text prompt should then focus almost entirely on motion. Describing the image again can work against you: the model may reduce movement or drift away from the original frame.

The image-to-video formula is short:

[subject motion] + [camera motion] + [scene/ambient motion] + [pace]

Examples you can paste under any start frame:

41. She turns her head toward the camera and smiles, hair moving slightly in the breeze; slow dolly-in; leaves drift in the background; calm pace.
42. The car drives forward out of frame; camera stays locked-off; dust rises from the road and settles; real-time speed.

Variations that work on most photos:

• "The subject blinks and takes a slow breath; subtle handheld sway; steam rises from the cup."
• "The product rotates 90 degrees on the table; camera holds still; light sweep across the surface."

For faces and products that must stay exact, generate or pick the start frame first, then animate it. That is usually more reliable than asking a text-to-video model to invent the same face twice. Kenerate's image to video tool takes a start frame on most models, and several also accept an end frame.

How do you prompt sound and dialogue in AI video?

Models with native audio, such as Veo 3.1, Seedance 2.x and Wan 3.0, generate sound from the same prompt. Google's Veo guide uses three conventions that also carry over well to other sound models:

• Dialogue in quotes: A woman whispers, "We have to leave now."
• Sound effects labelled: SFX: thunder cracks in the distance.
• Ambience labelled: Ambient noise: the quiet hum of a spaceship bridge.

If you plan to add your own music later, say "no music" in the prompt. On Kenerate the sound-capable models also have an audio toggle, and turning audio off lowers the price on some of them (Veo 3.1 Fast is 45 credits per second with sound and 30 without, per our pricing config). For more dialogue detail, see our Veo 3 prompt guide.

Which AI video model should you use for these prompts?

The prompts above are model-agnostic, but models differ in length, resolution and price. These figures come from Kenerate's video model config as of 18 October 2026 (credits for a 5-second clip at 720p, with sound where the model supports it; Veo uses its 4/6/8-second options).

A practical workflow: test a prompt on a cheap model at 480p or 720p, fix the wording, then render the keeper on the model and resolution you actually need. New accounts get 25 free credits to try the AI video generator; video generation itself requires a paid plan.

How to use these AI video prompts step by step

1. Pick the mode. Text-to-video for a scene from scratch (text to video); image-to-video when you need an exact face, product or composition.
2. Copy a prompt from the category closest to your idea and swap the subject, setting and style words.
3. Set the aspect ratio in the app, not only in the prompt: 16:9 for YouTube, 9:16 for Reels/TikTok, 1:1 or 4:5 for feeds.
4. Choose a short length first (4–5 seconds). Short clips follow prompts more closely and cost less.
5. Generate, then change one variable at a time: camera move, then lighting, then action. Note what changed.
6. Use a prompt enhancer if you're stuck. Kenerate's prompt enhancer expands a short idea into a structured prompt you can edit.
7. Render the final version at the resolution you need, or upscale the best draft.

What are the most common AI video prompt mistakes?

• Describing a photo instead of a shot. No action, no camera, so you get a near-static clip.
• Conflicting moods. "Dark horror, bright cheerful colours" splits the model's attention (Adobe lists this as a common error).
• Too many events. Three actions in 5 seconds usually means none of them reads clearly.
• Long on-screen text. Most models still garble more than a few words of text; add titles in an editor instead.
• Rewriting everything after one bad result. Fix the one part that failed and keep what worked.
• Forgetting audio. On sound models, unspecified audio may include music you don't want. Say "no music" or describe the sound.

Sources

• Google Cloud: Ultimate prompting guide for Veo 3.1
• Adobe Firefly: AI video prompts guide
• HeyGen: How to write AI video prompts
• Runway Help Center: Image-to-video prompting