Glacier lake
Aerial shot gliding over a turquoise glacier lake surrounded by snowy peaks, sunlight glinting off the water. Slow forward movement. Wind and distant water.
AI Video Pro · Text to Video
Text to Video AI: Turn a Sentence Into a Video
No footage, no photo. Describe the scene, the action, the camera and the sound, pick a shape and a length, and AI Video Pro renders it, with audio, in minutes.
Examples
Seaside morning
A pastel seaside town at morning, fishing boats returning to the harbor, gulls circling, a baker opening his shop shutters. Slow pan. Waves, gulls, shutter rattle.
Lantern festival
Hundreds of paper lanterns rise into a night sky over a river, reflections shimmering. Slow tilt up. Soft crowd murmur and water.
Foggy pier
Close-up of a man in a wool coat on a foggy pier, turning his head toward the sea as a foghorn sounds. Slow push in. Waves and a distant foghorn.
Mountain train
A red train winds through snowy mountains, steam trailing, the camera flying alongside. Train rhythm and wind.
City at dusk, spoken line
A woman in a camel coat walks toward the camera along a quiet city street at dusk and says, "I almost didn’t come back." The camera tracks backward at her pace. Warm shop windows, light traffic sound, her footsteps on wet pavement.
Text to video turns a written description into a moving shot. In AI Video Pro you leave the image slot empty, type the prompt and choose the engine: the newest engines hold a single take for up to thirty seconds and generate ambience, effects and spoken lines in the same pass.
A prompt that reads like a shot description gets a shot. The twelve prompts below follow the same five-part structure (subject, action, camera, light, sound) and cover the things people make most: product spots, travel scenes, dialogue, nature and stylised animation.
How it works
- 1
Write the scene
Describe who or what is in the shot, what happens, how the camera moves, the light and two or three sounds. Or copy a prompt below.
- 2
Pick shape, length and engine
Choose 16:9, 9:16 or 1:1, a length up to 30 seconds and an engine. The credit cost updates as you change them.
- 3
Generate and refine
Watch the result, change one thing in the prompt and run it again, or extend the clip from its last frame.
What people make with it
Product spots
A turntable, a pour, an unboxing: product clips with sound, without a studio.
Travel and places
Establishing shots of cities, coasts and landscapes for intros and b-roll.
Dialogue scenes
Put a line in quotes and the engine generates the speech, lip-matched to the speaker.
Stylised animation
Anime, watercolour, clay or pixel-art scenes from a single paragraph.
B-roll for edits
Fill gaps in a video essay or a podcast cut with matching shots.
Storyboards in motion
Turn a script beat into a moving previz before filming starts.
Prompt library
12 prompts distilled from the generations that get finished and downloaded. Copy one, or open it in the generator with the photo you want to animate.
City at dusk, spoken line
A woman in a camel coat walks toward the camera along a quiet city street at dusk and says, "I almost didn’t come back." The camera tracks backward at her pace. Warm shop windows, light traffic sound, her footsteps on wet pavement.
Lines in quotes are spoken and lip-matched.
Seaside morning
A pastel seaside town at morning, fishing boats returning to the harbor, gulls circling, a baker opening his shop shutters. Slow pan. Waves, gulls, shutter rattle.
Glacier lake
Aerial shot gliding over a turquoise glacier lake surrounded by snowy peaks, sunlight glinting off the water. Slow forward movement. Wind and distant water.
Lantern festival
Hundreds of paper lanterns rise into a night sky over a river, reflections shimmering. Slow tilt up. Soft crowd murmur and water.
Product turntable
A matte black wireless speaker on a white plinth rotates slowly through a full turn. Soft key light from the left, a thin rim light traces its edge. The camera stays locked off at eye level. Subtle room tone, no music.
Chef at the pass
Close-up of a chef’s hands plating a dish: a swipe of sauce, three herbs placed with tweezers, a final drizzle of oil. Shallow depth of field, warm kitchen light. Sizzle and clatter from the kitchen behind.
Honey close-up
Macro shot of honey drizzling onto a stack of pancakes, butter melting. Warm morning light, camera still. A soft sizzle.
Anime rooftop
Anime style. A young woman in a long scarf stands on a rooftop at sunset, the city glowing below. Wind lifts the scarf, petals drift past, the camera orbits slowly around her. Soft wind and a distant train horn.
Mountain train
A red train winds through snowy mountains, steam trailing, the camera flying alongside. Train rhythm and wind.
Day to night skyline
A locked-off wide shot of a city skyline. The sun sets, the sky turns orange then deep blue, and the windows light up one by one until it is night. Rising evening city ambience.
Try 30 seconds on an engine that supports it.
Foggy pier
Close-up of a man in a wool coat on a foggy pier, turning his head toward the sea as a foghorn sounds. Slow push in. Waves and a distant foghorn.
Watercolour village
Watercolor animation style. A quiet hillside village in the rain, lanterns glowing, a woman with a red umbrella crossing a stone bridge. Slow pan. Rain and a distant bell.
Write it like a shot, not a wish
Engines follow direction. "A beautiful video of the ocean" leaves every decision to chance; "low angle at the waterline as a wave rolls in and pulls back, locked-off camera, surf and a gull" makes the decisions for them. Name the camera move, the light and the sounds, and the result stops being random.
Give the action as many beats as the clip is long. One beat fills five seconds; a thirty-second take needs three or four events in order.
Which engine to pick
Every engine in AI Video Pro accepts text. The engine guides on the AI Video Model Index compare length, resolution, audio, live render time and price per clip, so you can pick the right one for each shot and switch without changing the prompt.
Frequently asked questions
- How does text to video AI work?
- You describe the scene in words and a video model renders it frame by frame, with sound on engines that support native audio. In AI Video Pro, leave the image slot empty and generate.
- How long can a text-to-video clip be?
- Up to 30 seconds in a single take on the longest engines, and 5 to 15 seconds on the others. Any clip can be extended afterwards.
- Does it make sound?
- Yes, on engines with native audio. Name the sounds in the prompt, and put spoken lines in quotes.
- Which aspect ratios are available?
- 16:9 for widescreen, 9:16 for phone-first vertical video and 1:1 for square posts, with 4:3 and 3:4 on some engines.
- What does it cost?
- Text to video is paid with credits or a subscription. The credit cost for the engine, length and resolution you pick is shown on the Generate button before you confirm.
- Are there content rules?
- Yes. GoCrazyAI is for adults (18+). Prompts about real people or that break the community guidelines are declined automatically.
Ready to try it?
Open AI Video Pro, upload a photo and paste a prompt from this page.
Last updated: