Animate one existing visual. Use this mode when you have an approved key visual and want to bring it into motion.
1
Open a Video Playground
From the Playgrounds page, create a new Video Playground. You can also send any image directly to a new or existing Video Playground from the Gallery or from an image playground.
2
Select "Single image" mode
Single image is the default mode. Upload your image by selecting the image area or dropping a file in.
3
Write a prompt
Describe how you want the image to move. A text model can help write a clear prompt.
4
Add reference images (optional)
Add up to 3 reference images to guide the visual style of the output. References work the same way as in the image playground.
5
Set aspect ratio and duration
Choose an aspect ratio: Landscape (16:9) or Portrait (9:16).
Choose a duration: 4 seconds (default), 6 seconds, or 8 seconds.
Some models may not support all settings. When that's the case, a note appears below the control: "Some models may ignore this setting."
6
Generate
Select Generate. You will be taken to the canvas. Each model generates one version of your video, and versions appear on the canvas as they finish.
First to last frame
Control exactly how the video starts, ends, or both. Use this mode when you know the opening and closing frames and want the model to generate the motion between them.
1
Open a Video Playground
From the Playgrounds page, create a new Video Playground.
2
Select "First to last frame" mode
Select First to last frame in the mode selector.
Upload a start image, an end image, or both. The model generates the transition between them.
If you already uploaded an image in Single image mode and then switch to First to last frame, that image becomes the start frame automatically.
3
Write a prompt
Describe the motion you want between the two frames.
4
Set aspect ratio and duration
Choose an aspect ratio: Landscape (16:9) or Portrait (9:16).
Choose a duration: 4 seconds (default), 6 seconds, or 8 seconds.
5
Generate
Select Generate. Versions appear on the canvas as each model finishes.
Reference images and the Loop option are not available in First to last frame mode.
Review and export
Each generation produces multiple versions at the same time. Results appear on the canvas as Model A, Model B, and so on. The model behind each label is not shown, so your choice stays focused on the output itself.
1
Browse your versions on the canvas
Hover over any video card to play it. All versions from one generation appear on the same row, connected to their source scene.
2
Open a video to view it
Select the eye icon on any video card to enter the detail view. The video plays on the left. Its settings (prompt, frames used, ratio, duration) are visible on the right.
From this view, you can switch between versions without going back to the canvas.
3
Vote
Select ๐ or ๐ on the version you are reviewing. You can vote from the canvas or from the detail view.
Voting is required before you can export. Your votes also help TheSEA understand which models perform best for different types of content, and directly shape what gets built next.
4
Export
Once you have voted, the Export button becomes active. Select it to download the video as an MP4.
FAQ
What does the "Early access" tag in the header mean?
Video is still in pilot. Results may vary as the models and the product continue to improve. Your votes and feedback directly shape what comes next.
How do I write a video prompt?
Start with your idea and the image you want to animate. Share both with a text model, such as Gemini or Claude.
[Describe what you want to happen in plain language].
The product is [brief description โ color, material, shape].
The mood should feel [one word or reference].
The product should [move / not move].
The camera should [stay still / move โ describe roughly how].
Write me a video generation prompt under 2000 characters.
How do I get access to Video Playground?
Video is in early access and enabled at the organization level. Contact your account manager to request access for your organization.
Why is the Export button disabled?
You need to vote (๐ or ๐) on the version you want to export before the button becomes active. Select your vote, then export.
One version shows "Generation blocked." What should I do?
Some models reject specific inputs due to content policies. Select Adjust on the blocked card. You will return to the creation view with all your original settings pre-filled. Adjust your prompt or input image and generate again.
One version failed. Can I retry?
Yes. Select the retry button on the failed card. If the same version keeps failing, try adjusting your input.
Can I generate with only one model?
Yes. Select the dropdown arrow next to the Generate button to open the model panel. All models are selected by default. Uncheck any you want to exclude. Your selection is remembered for the rest of the session.
Can I use reference images in First to last frame mode?
No. Reference images are only available in Single image mode.
What aspect ratios and durations are supported?
Aspect ratio: Landscape (16:9) and Portrait (9:16).
Duration: 4 seconds (default), 6 seconds, and 8 seconds.
A glossy drop of blue nail polish falls vertically in slow motion toward a blue surface in front of a Plage nail polish bottle.
The drop lands near the bottle, triggering a dynamic splash of thick blue liquid that curls upward around the bottle. The bottle remains completely dry and untouched.
The splash ripples and slowly settles back into a sleek, reflective liquid pool. High-end cosmetic product aesthetic, vibrant cobalt blue studio background, smooth slow-motion physics, lighting highlights on gold cap and glossy textures.
The scene opens overhead, directly above the Desiaqua bottle: a pure top-down view of the matte black cap, centered in frame, surrounded by fine golden sand sculpted into concentric ripple rings. Warm amber light rakes across the texture. Everything is still.
The camera begins a slow, fluid descent and rotation โ unhurried, cinematic โ tilting from overhead to a low three-quarter side angle, gradually revealing the full bottle: clear glass filled with deep amber liquid, the Desiaqua Aqua Oasis lettering catching the warm light, the bottle resting at a slight angle nestled in golden sand dunes.
The bottle never moves. Only the world around it breathes. As the camera moves, a soft desert wind stirs โ fine sand particles lift and drift lazily across the surface, catching the backlight as golden dust. The light shifts subtly, a warm golden ray sweeping slowly across the dunes, deepening shadows on one side, igniting the amber liquid from within.
The movement is barely perceptible โ a whisper of wind, a breath of light. The sand settles as the camera reaches its final position, holding on the full product side view.
Camera: smooth orbital move from 90ยฐ overhead to approximately 20ยฐ elevation side angle. No shake, no cuts.
Speed: extremely slow. Full move takes 8โ10 seconds of real time.
Lighting: warm golden hour. Strong directional side light. Soft glowing backlight through the amber liquid.
Mood: desert luxury, meditative, cinematic. Silence and stillness with the gentlest signs of life.
