Lemo-Opuscar

99 film styles for your coding agent. Pick one, bring your story, and the agent makes the film, entirely in code.

▶ Open the gallery · 中文画廊

Official repo: github.com/lemomo-ai/lemo-opuscar · by Lemomo (@lemomo_ai)

[!IMPORTANT] This release was audited and improved with Claude Fable 5.2.

Built for Claude. Claude Fable, Opus and Sonnet all work; set effort to high (in Claude Code, type /effort high). Other models may not reproduce the styles well, so use them with care.

🆕 New

• 2026-11-12 · 33 new styles: 99 now, in 15 categories sorted by what the picture looks like, and the gallery in English and Chinese, with a page for every style. Some demo films remade, bugs fixed, style prompts improved. See the styles ↓
• 2026-10-22 · Make a style your own: change a style or start blank, and keep it. See how ↓

How it works

1. Pick a style. Browse the gallery by what the picture looks like, or search for what you are making: explainers, launches, news, music videos.
2. Watch its demo. Every style has a short film made in code, and a page with stills, what it suits and the STYLE.md your agent follows.
3. Ask your agent. Install the skill once, then say it in one sentence:

Make a film in the Neon Tubes style about the coffee farm my family runs.

Name the style in English or Chinese, or by its folder; the style index lists them all. Add a length, a narrator, your own photos or music if you like:

Make a 45-second film in the watercolor style about the coffee farm my family runs. Warm female narrator.

Make a 40-second LiDAR Survey film about the old quarry in the hills behind our town, from public laser-scan data.

Install

The skill (recommended). In your terminal:

claude plugin marketplace add lemomo-ai/lemo-opuscar
claude plugin install lemo-opuscar@lemolab

Then use it from any folder. On first use it downloads the guides, tools and style prompts to ~/lemo-opuscar, shared by all your films; each film's project, from source to finished video, goes in the folder you started from. For other agents, copy plugin/skills/lemo-opuscar/ into their skills folder.

Or clone the repo. Films go into films/<name>/ inside it.

git clone https://github.com/lemomo-ai/lemo-opuscar.git
cd lemo-opuscar
claude

Update or uninstall

Update, then restart Claude Code (the library in ~/lemo-opuscar updates itself on the next film):

claude plugin marketplace update lemolab && claude plugin update lemo-opuscar@lemolab

To have it update by itself: in Claude Code, type /plugin, open Marketplaces, pick lemolab and choose Enable auto-update (third-party marketplaces such as lemolab don't update automatically until you turn this on).

From a clone: git pull, then run sh plugin/skills/lemo-opuscar/scripts/setup.sh deps once (it updates only the core packages that changed; if you installed the voice or music packages, run … deps voice music instead).

After an update, ask for a film as usual; to make a style your own, see below.

Uninstall, then delete ~/lemo-opuscar:

claude plugin uninstall lemo-opuscar@lemolab

Your own styles in ~/lemo-my-styles/ stay through updates and uninstalls; delete that folder yourself if you don't want them. If a step stays stuck, open an issue.

What happens next

Not sure yet? With the skill installed, type /lemo-opuscar:lemo-opuscar and nothing else. It asks which language to use, then what you'd like to do: a film in a library style, make a style your own, or a film in a style you've saved.

It asks you once, up front: anything it can't decide about your topic; whether you have your own voice, music or other material; whether you want to see a storyboard first (say yes and it stops once to show you the key shots in the real style; otherwise it goes straight to the finished film).

Then the agent reads three guides and works like a small studio:

File	What it gives the agent
DIRECTOR.md	how to direct: story, sound, rhythm, camera, performance, self-checks
TECHNIQUE.md	how to build: frame-by-frame rendering, voice, music, mixing
styles/<style>/STYLE.md	what the style looks and sounds like; the story is yours

Every film is directed, drawn, scored and mixed by the agent writing code: 🎨 canvas and WebGL pages rendered frame by frame, 🎵 original scores from free sample libraries, 🎙️ text-to-speech narration. No video generation, no stock footage.

Before you start

• A film takes an agent about 30–60 minutes and a fair amount of tokens.
• You need Node 20+, ffmpeg 7+ and Python 3.11+ (or uv); the agent installs the rest. Ubuntu 24.04 ships ffmpeg 6.1, which is too old: use a static build.
• Three styles draw their characters with an image generator you bring (Manga Panels, Text Flash Cuts, Comic History Cartoon; marked ✦ in the style list): the Codex CLI signed in to an account that can generate images, or another image tool you have. Without one, those three can't start; every other style works without it.
• Default output 1920×1080 at 24 fps (some styles set their own frame rate, such as 30 fps); other sizes on request.

This is the beginning of the README. Read the rest at the original. Read the full README ↗