VIDEO AGENT SHOWCASE

Code to video. Ideas in motion.

Explore editable motion graphics, SVG illustrations and Three.js scenes created with Dojo’s video agent and engine. Find a direction, copy the brief, then make it your own.

Three.jsSVGMotion graphicsMP4
Quiet Current

Quiet Current

SVG · 8s

01 — 06

Orbital Bloom
3D / Three.js8s · Silent

Orbital Bloom

Ceramic rings and satellites form a rotating 3D sculpture.

Quiet Current
SVG illustration8s · Silent

Quiet Current

Two koi circle through layered SVG water and ornamental linework.

Make It Move
Mixed motion10s · Silent

Make It Move

Bold typography and an assembling 3D cube become a short brand film.

Chromatic Tunnel
3D / Three.js8s · Silent

Chromatic Tunnel

Neon frames recede into a rotating optical tunnel.

Prism Tide
3D / Three.js8s · Silent

Prism Tide

A field of ceramic columns moves in a synchronized wave.

Pocket Observatory
3D / Three.js10s · Silent

Pocket Observatory

An orbiting camera explores a miniature observatory on a floating island.

Curated agent-generated examples. Some scenes needed follow-up edits or technical fixes. Prompts are reusable starting points; results can differ.

From this example to your idea.

Open video agent
01

Choose a direction

Start with an illustration, a 3D scene or a complete motion sequence.

02

Copy the brief

Change the subject, colors, words or format to suit your project.

03

Refine in the video agent

Ask for targeted changes, review the result and export your video.

About code-to-video

What does code-to-video mean?

The agent creates structured graphics and animation instructions. Dojo’s engine renders them into video, keeping the project’s supported text, shapes, timing and 3D objects editable.

Do I need to write code?

You can start with a natural-language brief. Describe what should appear and how it should move, then ask for changes. Developers can also work with Dojo’s open-source engine and harness.

Can GPT generate these videos?

A language model can author the scene instructions; the video engine renders the frames. These examples use that workflow rather than a model that generates photographic footage directly.

Can I paste any SVG or Three.js code?

The graphics workflow supports a bounded editable SVG subset and validated 3D scene data. It does not execute arbitrary JavaScript or promise support for every SVG, shader or external 3D model.

Will the same prompt produce exactly this result?

Not necessarily. These are reviewed examples, and some include refinement prompts. Use them as a creative starting point. Keep the project to edit its layers; an exported MP4 is a flattened video.

Looking for generated footage?Explore AI video examples