Lemo-Opuscar

// Overview

Lemo-Opuscar is a toolbox of film styles for coding agents, published by Lemomo (LemoLab) as a Claude Code plugin and skill. Each of its 59 styles is a style prompt (`STYLE.md`) plus a demo film made entirely in code, and you pick one, bring a story, and let the agent direct. The picture is canvas and WebGL pages rendered frame by frame, the music is an original score built from free sample libraries, and the voice is text-to-speech narration; the README stresses that no video generation model and no stock footage are involved.

The agent works from three guides: `DIRECTOR.md` covers story, sound, rhythm, camera, performance and self-checks, `TECHNIQUE.md` covers frame-by-frame rendering, voice, music and mixing, and each `styles/<style>/STYLE.md` describes what that style looks and sounds like while leaving the story to you. Before starting it asks once about anything it cannot decide, whether you have your own voice, music or material, and whether you want to see a storyboard of key shots first. The styles are grouped into families such as Edit & Documentary, Evidence Films (built from real footage, documents and data, every number sourced), Hand-drawn & Painting, East Asian Traditions, Print & Printmaking and Graphic & Type.

A "make a style your own" flow turns a library style into a personal one, or builds a new one from a description or a reference picture: the agent plans the changes one question at a time, renders a style frame beside the original, then makes a 15 to 30 second demo film, and kept styles live in `~/lemo-my-styles/` where updates never touch them. The README warns that the styles are tuned for Claude Opus 5.5 and other models may not reproduce them well, and that a film takes an agent about 30 to 60 minutes and a fair amount of tokens. Code and docs are MIT; the demo films and their stills are CC BY 4.0, apart from three voice tracks the README excludes.

// What it does

• 59 ready-made film styles, each a STYLE.md style prompt plus a demo film made in code, browsable in an online gallery
• Director and technique guides that make the agent plan story, camera, rhythm and sound and check its own work

• Picture, music and narration all produced by code: canvas and WebGL frames, scores from free sample libraries, and text-to-speech voices
• An optional storyboard stop that renders the key shots in the real style before the full film
• A make-it-your-own flow that adapts a library style or starts from a blank page, shows a side-by-side style frame and a demo film, and saves the result to ~/lemo-my-styles
• Installs as a Claude Code plugin; the skill folder can be copied into other agents' skills directories

// Getting started

You need Node 20+, ffmpeg and Python 3.11+ (or uv); the agent installs the rest. Default output is 1920×1080 at 24 fps, with other sizes on request.

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

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

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

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

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Turn a topic or story into a narrated short film in a distinct visual style without a video generation model
• Make data-driven explainers in the Evidence Films styles, built from real footage, documents and public data

• Build a channel or brand look of your own from a library style and reuse it for every later film

// How Lemo-Opuscar compares

Lemo-Opuscar alongside other open-source video generation tools AI/TLDR tracks, ranked by GitHub stars.