Here are the tools behind the animated videos, with a quick explanation of what each one does and links to try them.

The core tools

1. Claude Opus 5.5: the AI doing the work

Use Opus 5.5 in a coding agent such as Claude Code. It plans the scenes, writes the animation code, and uses the connected tools to create the video.

Get Claude

2. HyperFrames: turn code into video

HyperFrames is a free, open-source video framework. Your agent builds each video as an HTML page, then HyperFrames checks it for layout problems, captures preview frames for review, and renders it to MP4.

My workflow uses v0.8.76, run through npx.

Get HyperFrames

3. GSAP: animate the scenes

GSAP is the animation library inside the HTML. It controls streaming headlines, camera moves, stamps, wipes, and the timing of elements on screen.

Get GSAP


Images, voice, and music

4. GPT Image 2: create the illustrations

OpenAI's GPT Image 2 creates the stylized illustrations used in the videos. You can access it through Higgsfield.

For the examples shown, I use a consistent visual style with no people and no text in the generated images. Headlines and labels are added separately in the animation.

Try GPT Image 2 on Higgsfield

5. ElevenLabs: generate the voiceover

ElevenLabs creates the AI narration. My setup uses the premade stock voice Brian with the eleven_v3 model.