Creative · Design

huashu-motion turns a coding agent into a one-person film studio: it writes the film as code, composes the music, and takes frame-level notes from you.

A Claude Code skill that ships 39 film styles as prompts — and proves each one with a demo film its agent rendered, scored, and narrated entirely in code.

Agent-side workflows that turn novels into production-ready short-drama bibles—characters, outlines, and art guides—without leaving Claude Code or Codex.

Those viral Claude Opus 5.5 animations are just code the model wrote — this list collects them with the prompts that produced them, so you can copy the technique instead of guessing.

CozyClay puts previsualization in a single browser tab so you can stage scenes, pose characters, and hand the camera off to an AI model or an LLM agent.

OpenFilm turns Claude Code or Codex into a video agent by letting it write films as plain web pages, then hands you a real editor to fix what it made.

Mesh Avatar Studio has Claude Code or Codex do the tedious rigging work of turning one illustration into a live 2D avatar, then hands you a drag-to-fix editor.

Papermorph converts PDFs and stories into static, narrated, interactive web lessons — and paper-puppet films — with quizzes, progress, and no server in sight.

Agents can draw SVGs all day; what they lack is the unglamorous craft of identity design — the brief, the optical corrections, the 16-pixel test, and knowing when to stop for approval.

Oil UI makes coding agents produce several genuinely distinct UI directions, put them side by side, and refine against real screenshots.

Dioramas is a free three.js framework for cinematic one-object websites, with a prompt-to-model asset pipeline and 20 finished examples to prove it.

A Claude Code skill that packages motion-design grammar, audio rules, and a builder-vs-judge review loop so AI-generated business videos actually look shippable.

A curated archive of 18 copy-ready prompts and 54 real cases where Claude Opus 5.5 rendered entire videos by writing code, not pixels.

onetake is a Claude Agent Skill that makes product films in one unbroken take — and auto-rejects any video where one beat simply replaces the next.

A structured knowledge kit that forces Claude to design like a senior architect instead of an enthusiastic intern.

BeefTV is a local-first desktop workspace that treats AI as a collaborator on a free-form creative canvas rather than a button you click.

It exists because agents default to cliché unless you feed them the vocabulary of craft.

This agent skill bundles 123 scraped Apple HIG pages so your coding agent critiques UI with citations instead of vibes.

anidoodle turns illustration, animation and even film scores into deterministic TypeScript, so the same source redraws the same picture on every machine.

General image models produce logo-shaped pictures with nothing behind them; Inkloom is building a pipeline that constructs a mark as a sequence of explainable decisions — and says plainly that the generator is not live yet.