Repolode
Submit
Lob0Garou

Lob0Garou/opus-visual-motion-engine

An agent skill giving coding models a gated pipeline for motion graphics and code-rendered video: plan, render, automated QA, and MP4 export.

View on GitHub
Stars
3
Forks
0
Open issues
0
Last commit
26 days ago

What is it?

opus-visual-motion-engine is an agent skill for producing motion graphics, code-rendered video, kinetic typography, landing pages and diagrams. It walks a coding model through a fixed sequence: write a plan with beats and a timeline, build from a 1920×1080 HTML template that exposes a seek(t) and DURATION contract, render headlessly, run automated QA, fix issues, and export to MP4, WebM or GIF with ffmpeg.

The QA script renders each frame and checks for frozen or empty frames, non-deterministic output, text collisions, contrast and minimum type size, and common AI-generated design patterns. It writes a contact sheet, key frames, and a report.json with per-frame ink coverage, so a model without image input can still reason about the result. Page mode also checks desktop, mobile and dark-scheme layouts.

The repository also includes a score script that generates deterministic background music and sound effects from a cue sheet, and an audio-data script that extracts beat, onset and loudness data so motion can sync to sound.

Key features

  • 01Plan-to-export pipeline with a reusable 1920×1080 HTML composition template.
  • 02Deterministic seek(t) timing contract with a seeded random generator and easing helpers.
  • 03Headless render QA that flags frozen frames, text collisions, low contrast and scene gaps.
  • 04Exports frame-exact MP4, WebM or GIF via ffmpeg, with optional audio.
  • 05Generates original music and sound effects from a cue sheet, and extracts beat data for syncing motion to audio.

Use cases

  • Producing short motion graphics or showreels with a coding agent.
  • Building code-rendered video or kinetic typography without a manual editor.
  • Checking an agent-generated animation for layout and timing defects without seeing the frames.
  • Creating landing pages or diagrams with animated transitions that are verified by automated checks.

Quick start

  1. 1

    Clone the repository into your harness skills folder

    git clone https://github.com/Lob0Garou/opus-visual-motion-engine ~/.dsh/skills/opus-visual-motion-engine
  2. 2

    Install the script dependencies and the Chromium browser

    cd ~/.dsh/skills/opus-visual-motion-engine/scripts && npm install && npx playwright install chromium
  3. 3

    Start a new session and ask the agent to use the skill; then run the QA check on its composition

    node scripts/qa.mjs composition.html

Things to know

  • It is an agent skill for coding models, so it needs a compatible agent harness to be useful.
  • Export requires ffmpeg on PATH and Node 18 or newer.
  • The only benchmark is a single run per arm, so the reported gains are not statistically established.
  • The project is new, with only 3 stars.

Topics

#agent-skills#motion-graphics#video-generation#ffmpeg#animation#visual-qa#ai-agents#deepseek

FAQ

What is opus-visual-motion-engine?
An agent skill giving coding models a gated pipeline for motion graphics and code-rendered video: plan, render, automated QA, and MP4 export. opus-visual-motion-engine is an agent skill for producing motion graphics, code-rendered video, kinetic typography, landing pages and diagrams. It walks a coding model through a fixed sequence: write a plan with beats and a timeline, build from a 1920×1080 HTML template that exposes a seek(t) and DURATION contract, render headlessly, run automated QA, fix issues, and export to MP4, WebM or GIF with ffmpeg.
Is opus-visual-motion-engine free to use? What license does it have?
opus-visual-motion-engine is open source under the MIT license. Check the license terms before using it in commercial products.
Is opus-visual-motion-engine still maintained?
Actively maintained. The last commit to opus-visual-motion-engine was on Sep 26, 2026.
What can I use opus-visual-motion-engine for?
Producing short motion graphics or showreels with a coding agent. Building code-rendered video or kinetic typography without a manual editor. Checking an agent-generated animation for layout and timing defects without seeing the frames. Creating landing pages or diagrams with animated transitions that are verified by automated checks.
How do I install opus-visual-motion-engine?
Clone the repository into your harness skills folder: git clone https://github.com/Lob0Garou/opus-visual-motion-engine ~/.dsh/skills/opus-visual-motion-engine Install the script dependencies and the Chromium browser: cd ~/.dsh/skills/opus-visual-motion-engine/scripts && npm install && npx playwright install chromium Start a new session and ask the agent to use the skill; then run the QA check on its composition: node scripts/qa.mjs composition.html
Where can I get opus-visual-motion-engine?
The source code and installation instructions are on GitHub: https://github.com/Lob0Garou/opus-visual-motion-engine

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects