One prompt, code-only animation: generating a 40s LLM explainer with Remotion, SVG and open-source TTS
eyishazyer · x · 2026-09-25
The author shares the full single prompt behind a 40-second animated video explaining how LLMs work, where every frame is generated from code — no stock footage, no video generators, no image editors.
Technical requirements in the prompt:
- Remotion (React/TypeScript) for animation; all visuals drawn in SVG and Canvas
- Open-source TTS for narration; a light playful score synthesized in Python
- Export MP4 at 1080x1080 (square for X), 40 seconds total
The video is structured by section: 0-5s hook (a glowing dot on black, typewriter text "Ever wonder how AI actually thinks?", the dot becomes a cute wide-eyed blob), 5-12s tokens as input, and so on. The author says they're hooked and generates a new video hourly with varied prompts.
Related event: Claude renders a 40-second animation with code, no video model needed(2 posts)→
More from coding & agent
- Building a dynamic model router with LangChain and Jev middleware — Arindam_1729 · 2026-09-25
- Claude Code users urged to run '/claude-api prompt-audit' to tune configs for Opus 5.5 — JeremyNguyenPhD · 2026-09-25
- Claude Opus 5.5 builds a playable Splatoon-style game demo — Angaisb_ · 2026-09-25
- Dev releases jev-route, a framework-agnostic tool router for agents built on TypeSafe's Jev decision model — Huzeyfe_dev · 2026-09-25
- Renaming an API field is the most dangerous breaking change — diff the spec in CI — Franc0Fernand0 · 2026-09-25
- One-click narration pipeline: Claude writes shot-aware script, ElevenLabs MCP voices it — nikola_mr64990 · 2026-09-25