lemo-opuscar: 66 film styles made in code — but Opus 5.5 only

Claim

66 film styles made in code?

Catch

TUNED FOR OPUS 5.5

Verdict

Not magic, OPUSCAR

"Sixty six film styles, each a demo film made entirely in code by Claude Opus 5.5 — no video model, no stock footage." That's the pitch behind lemo-opuscar, an agent skill for Claude Code from Lemomo. It's not an AI video generator — every frame is drawn by code. Here's what the repo actually says, and the catch.

The claim

lemo-opuscar is "a toolbox of film styles made in code for Claude Code: 66 ready to use, and your own." The GitHub API puts the repo at 9,308 stars and 1,040 forks, created 26 September 2026, under the MIT licence. The gallery lists 66 styles across 11 categories, totalling 63 minutes of film. Every style is a prompt plus a demo film, directed, drawn and scored entirely in code by an AI agent — Claude Opus 5.5. No video generation, no stock footage.

The flagship film is OPUSCAR 98: all 98 Best Picture winners from 1927 to 2025 in 6 minutes 31 seconds, each redrawn in a fitting style. The 98 styles break down as 77 in 2D, 4 in 2.5D and 17 in 3D.

What it actually produces

You install the skill with claude plugin marketplace add lemomo-ai/lemo-opuscar and claude plugin install lemo-opuscar@lemolab, then describe the film you want. You need Node 20+, ffmpeg 7+ and Python 3.11+ (or uv); the agent installs the rest. One film takes an agent about 30–60 minutes and a fair amount of tokens, and defaults to 1920×1080 at 24 fps. Films you make are yours, no credit needed; the library code and docs are MIT, and the demo films are CC BY 4.0.

The catch

• Opus 5.5 only. The styles are tuned for Claude Opus 5.5 — other models may not reproduce them well.
• Local toolchain. Node 20+, ffmpeg 7+ and Python 3.11+ are required. (Ubuntu 24.04 ships ffmpeg 6.1, which is too old; use a static build.)
• Time and tokens. A film takes roughly 30–60 minutes and a fair amount of tokens.
• Brand new. Created September 2026, with almost no independent reviews yet.

Try it

lemo-opuscar repo: https://github.com/lemomo-ai/lemo-opuscar Gallery: https://lemomo-ai.github.io/lemo-opuscar/ README (install & requirements): https://github.com/lemomo-ai/lemo-opuscar/blob/main/README.md

Verdict

lemo-opuscar is a genuine, code-drawn film library — not an AI video generator — and the numbers check out: 66 styles across 11 categories, 9,308 stars, OPUSCAR 98 covering 98 Best Picture winners in 6:31. The catch is the setup: Claude Opus 5.5 specifically, plus Node 20+, ffmpeg 7+ and Python 3.11+, and 30–60 minutes per film. Every number traces to the repo README, the gallery and the GitHub API.

Sources: github.com/lemomo-ai/lemo-opuscar · api.github.com/repos/lemomo-ai/lemo-opuscar · lemomo-ai.github.io/lemo-opuscar/ · github.com/lemomo-ai/lemo-opuscar/blob/main/README.md

SOURCES

SHARE THE CATCH

Seen this claim going around? Send the person who shared it the source.