AI Video Generation

Generate videos from text prompts, start/end frames, or reference image sets.

Upload a required start frame and an optional end frame to generate a motion transition.

How to write prompts for Start + End Frames

Start frame

End frame

Optional. Leave this empty to animate from a single image.

0/2500

10 credits / second + 10 per video

Available durations: 3–15 seconds

A little inspiration, in motion.

Explore the prompts and inputs behind these clips.

Explore video examples
At the edge of the Atlantic
Text to video
A visitor above the rooftops
Text to video
Silver, in the spotlight
Crimson in the mist
Above the waking valley
Text to video
Paper wings over the ocean
Silver pleats, cobalt light
A little light in the forest
Seoul summer evening
Text to video
Painting a flamingo
Wool, on demand
Tailored tiger stripes
A Dalmatian, dot by dot
Antlers in bloom
Pressing out the wrinkles
Moon diver
Catching a galaxy
Aquarium balloon
Window kite
Lunar waiting room
Star rain
Fishing for starlight
The weightless moment
A giant in the city
Peel into a new dimension
AI video generator

Generate short AI videos from prompts, guide frames, or reference images

Create video concepts with text prompts, first-and-last-frame guidance, or reference-image mode. Adjust aspect ratio, choose a model, and iterate on motion, pacing, and composition for ads, storyboards, and creator content.

Read More Guides

Popular use cases

Storyboard motion tests

Turn still concepts into short motion studies before committing to a full production or editing timeline.

Campaign and ad iterations

Generate multiple visual directions for product launches, performance ads, promos, and social storytelling.

Pre-visualization for creators

Test camera movement, subject behavior, and pacing from prompts or guide frames before final assembly.

Related reading

Write better start and end frame prompts

Improve motion transitions, timing cues, and scene continuity when using frame-guided generation.

Open article
Write better prompts with reference images

Guide style, composition, and subject consistency with stronger reference-image prompt patterns.

Open article

What the generator includes

Prompt-driven video generation with frame-guided and reference-image modes
Aspect ratio and seed controls for more repeatable iterations
Upload support for start frames, end frames, and reference images after sign-in
Task history, status polling, and downloadable outputs for authenticated users

FAQ

What is the difference between frame mode and reference-image mode?

Frame mode works best when you already know the opening and ending look of the shot. Reference-image mode is better when you want to guide style, subject consistency, and visual direction across the clip.

What makes a good AI video prompt?

Describe subject action, camera movement, scene details, mood, and timing. The more specific you are about motion and composition, the more usable the result tends to be.

How is AI video generation priced?

Seedance Standard: 8 credits/second. Fast: 10 credits/second. Kling: 10 credits/second + 10 per video. See the total before generating.