Bilal Labs

How to make videos with Claude Opus 5.5

Claude doesn't render video. It writes the code that does. I'm an indie iOS developer, and every recipe below ships a real MP4 I made this way, the exact prompt, what the agent did, render time and, where I could measure it, the token cost. Source for all of them is in the opus-video-recipes repo.

The App Store preview for my app Lovely, from one prompt. See the recipe.

The workflow in four steps

Claude Opus 5.5 facts that matter here

Model ID claude-opus-5-5, released September 22, 2026. $4 per million input tokens, $20 per million output, $0.20 per million cache reads. 1M-token context, 128K max output, adaptive thinking always on (default effort: medium). Sources: Anthropic's announcement and the models overview. Video work is agentic and re-reads the same files, so cache reads are most of the bill.

Recipes

Which framework?

HyperFrames vs Remotion vs Manim for Claude-made videos: language, license, agent support and what each is good at, from their docs.

FAQ

Can Claude Opus 5.5 make videos?

Not directly: it outputs text, not video. But it writes code very well, and frameworks like HyperFrames (HTML) and Remotion (React) turn code into MP4. The model writes the composition, the framework renders it.

Which framework should I use with Claude?

HyperFrames if you want plain HTML and agent skills built for it, Remotion if you already work in React or need its ecosystem, Manim for math and diagram animation. See the comparison page for details.

How much does a Claude-made video cost?

The one measured run here, an App Store preview plus screenshots, cost about $5 at Opus 5.5 API list prices, mostly cache reads. Rendering itself is local and free.

More AI dev guides: subagent examples for Claude Code and Cursor.