ASMRIFY
ASMRIFY
Start free

AI Video Prompt Generator

Build a text-to-video prompt shot by shot: subject, action, scene, camera, light, motion and sound.

Roll or pick each part of the shot and get a prompt written the way video models read it: one subject, one action, one continuous shot. It is built from 20 subjects with 40 actions, 9 scenes, 7 camera moves, 5 looks, 4 kinds of motion, 4 sound styles and 2 lengths: 206,080 prompts, or type your own subject. Tuned for 8-second Veo 3.1 Lite clips and the free 4-second preview.

Build your video prompt

Your prompt

375 characters

Static wide shot. A red vintage convertible cruises past the camera on a coastal road, at golden hour. Look: soft, natural light. Motion: smooth slow motion. Sound: an engine purr and tyre hum, natural and close; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

Generate this video opens Prompt to Video with this prompt filled in; you press Generate there (free account needed). The prompt is built in your browser from 206,080 combinations; no AI writes it.

Anatomy of a video prompt

The order matters less than having every part. A built prompt, line by line:

Camera
"Slow dolly-in." How the shot moves comes first, so the model frames the scene around it.
Subject + action + scene
"A barista pours a slow heart of latte art into a wide cup, in soft morning light from a window." One subject, one action, one place.
Look
"Look: warm, glowing tones." The light and colour of the shot.
Motion
"Motion: smooth slow motion." The speed of the action.
Sound
"Sound: the hiss of the steam wand, natural and close; no background music, no narration."
Length + format
"One continuous 8-second shot … Vertical 9:16, photorealistic, no text on screen, no logos."

Camera moves and light: a cheat sheet

Every option in the builder, and what it does to the clip.

  • Static wide shot

    The camera holds still and the action plays out in frame. The safest choice for a 4-second clip.

  • Slow dolly-in

    The camera glides toward the subject: focus, tension, a reveal.

  • Tracking shot that follows the action

    The camera moves with the subject at its speed. Good for anything that travels.

  • Low-angle shot looking up

    Makes the subject look taller and more dramatic.

  • Handheld close-up

    A little natural shake, close in. Feels personal and filmed on the spot.

  • Slow orbit around the subject

    The camera circles the subject. Good for objects, products and still poses.

  • Overhead shot from directly above

    Top-down. Good for food, desks, water and patterns.

  • Soft, natural light

    Even daylight with gentle shadows. Reads as real.

  • High contrast with deep shadows

    Moody and dramatic: night scenes, suspense.

  • Warm, glowing tones

    Golden and cosy: sunsets, kitchens, candlelight.

  • Cool, muted colours

    Calm and overcast: fog, rain, quiet mornings.

  • Bright, punchy colour

    Saturated and upbeat, made for the feed.

Sound cues that work

  • Name the source of the sound: "the hiss of the steam wand", "tyres on wet asphalt", "waves on pebbles".
  • Say what to leave out: "no background music, no narration". Add your own track afterwards with Add Music to Video.
  • One sound per action beats a list of five: the model mixes them into mud.
  • On paid plans Veo 3.1 Lite generates the sound together with the picture, so it lands on the action.

Mistakes that break AI video clips

  • Two actions in one clip

    "Pours the coffee, then drinks it" asks for two shots. An 8-second clip fits one action: make the second one its own clip.

  • Text and logos on screen

    Models misspell words and invent logos. Leave them out of the prompt and add your title afterwards with Add Text to Video.

  • Real people and brands

    Do not name celebrities, real people or trademarked products. Describe an original person or object instead.

  • A pile of style words

    "Cinematic, 8K, masterpiece, ultra detailed" adds little. Concrete nouns, a camera move and a light do the work.

  • A wall of text

    Two to five sentences is plenty. Very long prompts are slower to get right and can be refused.

12 example prompts to copy

Built with the generator above. Copy one, or open it in Prompt to Video.

  • Low-angle shot looking up. A surfer carves along the face of a wave, on a windy, overcast day. Look: cool, muted colours. Motion: smooth slow motion. Sound: crashing surf, crisp and detailed; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Overhead shot from directly above. A hummingbird hovers at a red flower, at golden hour. Look: bright, punchy colour. Motion: quick and energetic. Sound: a soft wing buzz, rich and cinematic; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Slow orbit around the subject. A chef slices a ripe tomato in one smooth stroke, in a cosy, warm interior. Look: soft, natural light. Motion: real-time speed. Sound: sizzling oil and a knife on wood, crisp and detailed; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Low-angle shot looking up. A pianist's hands play a slow chord on a grand piano, late at night, lit by a single lamp. Look: cool, muted colours. Motion: smooth slow motion. Sound: a soft piano chord, rich and cinematic; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Tracking shot that follows the action. A chef tosses sliced vegetables in a flaming wok, in a bright, minimal white room. Look: soft, natural light. Motion: quick and energetic. Sound: sizzling oil and a knife on wood, natural and close; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Handheld close-up. An old fisherman hauls a rope onto the deck, on a bright, clear afternoon. Look: high contrast with deep shadows. Motion: real-time speed. Sound: lapping water and distant gulls, quiet and soft; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Slow dolly-in. A hot-air balloon drifts low over a winding river, on a foggy morning. Look: cool, muted colours. Motion: one unhurried, continuous move. Sound: the roar of the burner, rich and cinematic; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Static wide shot. A red fox trots across a frosty meadow, at blue hour, just after sunset. Look: high contrast with deep shadows. Motion: smooth slow motion. Sound: crunching frost and birdsong, natural and close; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Slow dolly-in. An old fisherman hauls a rope onto the deck, on a bright, clear afternoon. Look: warm, glowing tones. Motion: real-time speed. Sound: lapping water and distant gulls, crisp and detailed; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Tracking shot that follows the action. A lone hiker crosses a narrow wooden bridge, at blue hour, just after sunset. Look: bright, punchy colour. Motion: smooth slow motion. Sound: wind and gravel underfoot, quiet and soft; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Static wide shot. An old typewriter rings its bell as the carriage slides back, in soft morning light from a window. Look: soft, natural light. Motion: one unhurried, continuous move. Sound: clacking keys and a bright bell, rich and cinematic; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

  • Slow dolly-in. A chef slices a ripe tomato in one smooth stroke, in a bright, minimal white room. Look: warm, glowing tones. Motion: real-time speed. Sound: sizzling oil and a knife on wood, natural and close; no background music, no narration. One continuous 8-second shot, a single action from start to finish. Vertical 9:16, photorealistic, no text on screen, no logos.

Questions

What makes a good AI video prompt?

One subject doing one action in one continuous shot, plus a camera move, a light, a sound and a length. Name things concretely ("a barista pours latte art into a wide cup") instead of piling up adjectives ("amazing cinematic coffee video").

Does this tool use AI to write the prompt?

No. It builds the prompt in your browser from hand-written lists of subjects, actions, scenes, camera moves, light, motion and sound: 206,080 combinations, plus anything you type yourself. The AI part is the video model that renders it.

Can I use the prompt in other video generators?

Yes. It is plain English, so it works in most text-to-video models. It is written for short vertical clips; for a widescreen video, change the "Vertical 9:16" line.

How do I get sound in the video?

Describe it in the Sound line. On ASMRIFY paid plans, Veo 3.1 Lite generates the sound together with the picture, so the pour, the engine or the waves come out of the same render. You can add your own music track afterwards with Add Music to Video.

How long should a prompt be for an 8-second clip?

Two to five sentences: the prompts this builder makes run about 300 to 460 characters. An 8-second clip fits one action, so more detail rarely helps, and very long prompts can be refused.

Is the AI video prompt generator free?

Building and copying prompts is free and happens in your browser. Rendering on ASMRIFY needs a free account: a 4-second 480p preview on Grok 1.5 while the weekly free pool lasts, or 8 seconds at 720p with sound on Veo 3.1 Lite with a paid plan.

Guides

Make the clip

Turn the prompt into a clip

Start free with an account: a 4-second preview while the weekly free pool lasts. Paid plans render 8 seconds at 720p with sound on Veo 3.1 Lite.

Open Prompt to Video All AI video tools