Text to video AI turns a written prompt into a short video clip. Describe what is in the shot, what happens, how the camera moves and how it looks, and Seedance 1.5 Pro renders it, usually in about two minutes. Free to start.
About 2 minutes
Typical clip, measured
5 seconds
Per clip from this page
Words only
No photo needed
Each clip came from the prompt on its card, written once and rendered with Seedance 1.5 Pro at 5 seconds. Press Try to put a prompt in the box and change it.
Top-down crane shot over a busy city crossing in heavy rain: a sea of black umbrellas flows across the zebra stripes, and one bright red umbrella moves against the crowd. The camera starts high and slowly descends toward the red umbrella. Wet asphalt reflects the traffic lights, cinematic, moody blue tones.
FPV drone shot: the camera races low over a jungle river, climbs the face of a tall waterfall, then tips over the edge and dives down along the falling water into the mist. Fast, smooth, wide lens, bright tropical daylight.
A swarm of glowing jellyfish drifts upward through the dark deep ocean, their bodies pulsing with soft blue and pink light, long tentacles trailing. The camera rises slowly with them. Tiny particles float in the water, deep black background.
A red fox in a snowy meadow listens, crouches, then leaps high and dives headfirst into the deep snow, powder exploding around it. Slow motion, side tracking shot at the fox's height, low winter sun behind it, sparkling snow in the air, nature documentary look.
A glass chess king stands alone on a wooden board, then shatters into hundreds of pieces in extreme slow motion, shards spinning through the air and catching the light. Low angle close-up, dark room, one hard spotlight from above.
Extreme macro, underwater, inside a clear water tank seen from the side: drops of violet, gold and teal ink sink into the water and bloom into slow curling clouds that twist around each other as they fall. Static camera, black background, soft side light, every swirl sharp.
Stop-motion claymation: a small orange clay octopus with a round head and eight curly tentacles sits behind a tiny drum kit and drums a fast beat with four tentacles at once, the cymbals wobbling. Handmade look with visible fingerprints in the clay, slightly choppy frame rate, warm desk lamp light, static camera at eye level.
Anime style, hand-painted look: hundreds of paper lanterns float down a calm river at dusk under a stone bridge, their glow reflected in the water, cherry petals drifting through the air. The camera slowly pans along the river. Soft pastel sky, gentle and calm.
Text to video AI is a video model that makes a clip from words alone: you describe the shot, and it draws every frame, the movement and the light. No photo, footage or editing app is needed. Here it runs on Seedance 1.5 Pro, the fast Seedance version that works from text.
It runs in the browser, inside the Studio of the free AI video generator. Video needs a free account, and finished clips land in Generations, where you can watch them and share them by link.
Name four things in plain words: what is in the shot and what it does, where the camera is, how fast things move, and the look. One action fits a 5 second clip best.
One subject doing one thing: "a red fox crouches, then dives into deep snow". Three actions in one clip get rushed or dropped.
Say where the camera is and how it moves: crane down, tracking shot, push-in, orbit, FPV, static. A static camera keeps fine detail sharp.
Words like slow motion, drifts, races or wobbles set the pace. Slow motion stretches a split second, a splash or a shatter, over the whole clip.
End with the look, anime, claymation or documentary, and the light: low winter sun, one hard spotlight, dusk. They change a clip more than adjectives do.
We rendered the eight clips on this page on 9 Oct 2026 on Seedance 1.5 Pro at 5 seconds. Each was ready 1 min 31 s to 6 min 08 s after the send. The prompts ran from 39 to 55 words.
| Clip | Technique | Words in the prompt | Ready in |
|---|---|---|---|
| 01Crane down to the red umbrella | Crane shot | 51 | 1 min 59 s |
| 02FPV dive down a waterfall | FPV drone | 40 | 1 min 38 s |
| 03Rising with the jellyfish | Rising camera | 39 | 1 min 34 s |
| 04Fox dive in slow motion | Slow motion, tracking | 45 | 6 min 08 s |
| 05Glass king shatters | Slow motion close-up | 39 | 2 min 41 s |
| 06Ink blooming in water | Macro, static camera | 47 | 3 min 48 s |
| 07Clay octopus on drums | Claymation | 55 | 1 min 31 s |
| 08Lantern river, anime look | Anime style, pan | 44 | 1 min 39 s |
Three steps, all on this page.
Subject, action, camera and look, in one to three sentences. Or start from a technique under the box.
Sign in when asked. Seedance 1.5 Pro renders the clip, usually in about two minutes.
The clip lands in Generations. Change the camera or the light in the prompt and send again for a new take.
What people ask before their first clip.
Yes, it is free to start, with no subscription. Video uses more energy than pictures; energy refills when you share, post or invite friends, and a daily bonus tops it up.
Type a prompt in the box above: what is in the shot, what happens, how the camera moves and the look. Send it, sign in, and the clip arrives in Generations, usually in about two minutes.
One to three sentences. The prompts on this page run from 39 to 55 words. A longer prompt does not make a longer clip: 5 seconds holds one subject and one action.
Push-in, pull-back, pan, tilt, orbit, crane, tracking shot, FPV and a static camera all work in plain words. Name one move per clip.
Seedance 1.5 Pro makes 5, 10 or 12 second clips. This page sends 5 seconds; pick another length in the Studio settings in the chat.
The model fills in whatever the prompt leaves open. Name the camera and the light, keep to one action, and send again: every send is a new take.
Yes, with image to video: the photo becomes the first frame and the prompt only has to describe the motion.
No. It runs in the browser on a phone or a computer. Video needs a free account.
Reviewed by Arthur Down, editor · Updated · How we check facts
