video maker

Stop struggling with manual video editing and expensive SaaS subscriptions for product demos. This skill generates animated promo videos, explainers, and feature announcements directly from code. It produces high-quality MP4 files complete with synthesized voice-overs, generative music, captions, and thumbnails using a local, open-source engine.

What it does

• Multi-format rendering outputs videos in landscape (16:9), square (1:1), or vertical (9:16) for YouTube, LinkedIn, and TikTok.
• Narrated voice-overs uses the Kokoro model to generate clear speech from script text with custom pronunciation overrides.
• Dynamic motion graphics builds scenes as pure functions of time using a dedicated API for consistent, frame-perfect animation.
• Automated soundtracking generates background music and sound effects (pops, clicks, impacts) synced to specific script events.
• Compliance checking validates technical specs like loudness (-14 LUFS), text readability, and claim sourcing before delivery.

How it works

1. Research and script by analyzing your project code and branding to write a script and map out design tokens in brand.css.
2. Project initialization creates the folder structure, generates audio files via tts.mjs, and calculates the scene timeline with plan.mjs.
3. Scene development uses index.html and a stage API to build visuals that react to audio cues and frame timing.
4. Build and render runs the Playwright-based engine to capture frames and mix the final MP4, then performs a quality audit.

Frameworks & tools

The engine uses Node.js, Playwright (Chromium) for rendering, Kokoro for TTS, and FFmpeg for media encoding. All assets are generated locally without cloud dependencies.

Why this beats prompting it yourself

Generic AI video tools often hallucinate product features or ignore brand colors. This skill enforces strict adherence to your source code facts, ensures animations are deterministic, and automatically flags accessibility issues like tiny text or unsourced claims.

Use cases

• Generating a 30-second vertical teaser for a new feature launch on social media.
• Creating a 60-second landscape explainer for a technical product's landing page.
• Producing a square video walkthrough of an API integration for a developer newsletter.
• Batch-rendering promotional clips in multiple aspect ratios from a single brand definition.

Known limitations

Requires local Node.js environment and approximately 90MB for the initial voice model download. Complex 3D animations are not supported as scenes rely on HTML/CSS/Canvas logic.