website-to-hyperframes

Capture a website and produce a HyperFrames video from its design and content.

Updated Oct 16, 2025
One-click install
npx skills add https://github.com/hugotown/dotfiles --skill website-to-hyperframes-hugotown
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: website-to-hyperframes
Source: https://github.com/hugotown/dotfiles/tree/main/agents/skills/website-to-hyperframes
Command: npx skills add https://github.com/hugotown/dotfiles --skill website-to-hyperframes-hugotown

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Turning an existing website into a polished video (social ad, product tour, launch teaser) normally requires manual design extraction, scripting, storyboarding, voiceover, and animation work. This Skill automates that entire pipeline from a single URL. ## Core Features & Use Cases - Automated site capture: Runs npx hyperframes capture to extract screenshots, design tokens, fonts, assets, and animation metadata from any URL. - Seven-step gated workflow: Produces DESIGN.md, SCRIPT.md, STORYBOARD.md, TTS narration with word-level timestamps, per-beat HTML compositions, and validated output. - Deterministic rendering rules: Enforces GSAP timeline registration, asset cross-referencing, lint/validate/snapshot checks, and delivers a HyperFrames Studio preview URL. - Use Case: A user pastes a product URL and asks for a 15-second Instagram ad; the Skill captures the site's brand identity, writes narration and storyboard, generates voiceover, builds animated compositions, and returns a scrubbable Studio project link. ## Quick Start Ask the agent to capture https://your-site.com and turn it into a 15-second social ad video.

Frequently Asked Questions about website-to-hyperframes

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I turn a website into a video?▼

Provide a URL and the Skill runs npx hyperframes capture to extract screenshots, colors, fonts, and assets, then follows a seven-step pipeline: design reference, script, storyboard, voiceover, composition building, and validation. The result is a HyperFrames Studio project you can preview and render to MP4.

What video formats and durations does HyperFrames support?▼

It supports landscape 1920x1080, portrait 1080x1920 for Stories and TikTok, and square 1080x1080 for feeds. Typical durations range from 10-15 second social ads to 30-60 second product demos, with beat counts matched to the narration script.

Which TTS providers can generate the voiceover?▼

Three options are supported: Kokoro runs locally for free without an API key, ElevenLabs offers the widest voice selection via MCP tools, and HeyGen TTS returns word-level timestamps automatically. The Skill recommends auditioning two or three voices before committing.

Does the workflow render an MP4 automatically?▼

No. Rendering is on-demand only because it takes minutes per pass. The pipeline delivers a localhost HyperFrames Studio preview URL first, and only runs npx hyperframes render when the user explicitly asks for the final MP4.

Why do animations look correct in preview but broken in the rendered video?▼

The capture engine seeks non-linearly, so bare gsap.to() calls, stacked transform tweens on one element, and gsap.from() inside clip scenes break determinism. The Skill mandates attaching all tweens to the scene timeline, using fromTo, and hard-killing exits with tl.set().

Can I use Google Fonts or inline SVGs in the compositions?▼

No. Compositions must reference captured files via relative paths like ../capture/assets/ and use @font-face with the downloaded woff2 files. Inlining SVG data or linking Google Fonts is flagged as an anti-pattern during the asset cross-reference check.