A 13-minute narrated explainer with 72 HyperFrames scenes, built by Claude Code subagents

[lewislf]

Sharing a project that pushed HyperFrames further than I expected: a 13-minute explainer on neurotechnology where every narrated sentence is its own HyperFrames scene (72 in total), timed to the voice with hyperframes transcribe.

The scenes are generated from Python, rendered with hyperframes render, and checked with snapshot contact sheets before assembly. studio/GOTCHAS.md in the repo lists what broke along the way, which might help others doing long-form work.

Repo: https://github.com/lewislf/claude-code-video-studio

Youtube: https://www.youtube.com/watch?v=etgDV-RIp_I

reels-teaser-en.mp4