Mographify
blushdas/mographify
Briefing de conteúdo → motion graphics vertical 60fps em comando único
Ver no GitHubFerramenta que transforma um briefing de conteúdo em vídeo vertical animado (1080×1920, 60fps MP4) pronto para produção. Integra CLI determinístico com skill Claude Code para automação completa de content → HTML → MP4, eliminando necessidade de After Effects ou Premiere.
Instalação
npx convez-ai blushdas/mographify1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Produção em massa de Reels/Shorts a partir de estruturas de conteúdo
- Automação de vídeos verticais para social media com sistema de cores editorial
- Pipeline generativo que mapeia briefing → template visual → MP4
Veredito

Solução prática para creators e marketers que precisam de alta cadência de vídeos verticais. Arquitetura bem pensada (CLI determinístico + skill generativo) reduz fricção, mas exige Node.js + Playwright + ffmpeg na máquina.
README
A tradução para o português deste README ainda está na fila.
MOGRAPHIFY
Content brief → animated motion graphic → 1080×1920 60fps MP4
No After Effects. No Premiere. Just a content brief and a command.
MOGRAPHIFY v1.0.0
──────────────────
content → visual → 1080p 60fps MP4
◆ Input: claude-chat-vs-runtime.html (5 scenes × 10s)
◆ Duration: 52.0s total
◆ Output: ~/Desktop/
⟳ Recording 52s @ 60fps... done
⟳ Transcoding H.264... done
┌──────────────────────────────────────────┐
│ ~/Desktop/claude-chat-vs-runtime.mp4 │
│ 2.1MB · 52s · 60fps · 1080×1920 │
└──────────────────────────────────────────┘
What it does
Takes a structured content brief (HOOK / PREMISE / POINTS / WHY IT MATTERS) and produces a production-ready vertical video Reel — the kind you'd post on Instagram, TikTok, or YouTube Shorts.
The design system is strictly editorial:
- Black background, white text, one accent color per pillar
- Georgia italic for big statements, Helvetica Neue for labels
- Every scene holds 10s with animated entry and exit
- Accent element breathes (subtle opacity pulse) while on screen
Pillar-aware color system:
- Build / code / automation →
#FEF09Abutter yellow - Mindset / mental models →
#7DD3FCsky blue - Process / systems →
#86EFACmint green - Content / creation →
#FB923Cwarm orange - Strategy / positioning →
#C4B5FDsoft purple
Sample
samples/claude-chat-vs-runtime.html — 5-scene reel on Claude Code as a runtime vs chat interface.
Record it yourself:
mographify record samples/claude-chat-vs-runtime.html
Install
Prerequisites:
- Node.js 18+
- Playwright with Chromium
- ffmpeg
# Install prerequisites
npm install -g playwright
npx playwright install chromium
brew install ffmpeg # macOS
# Install mographify
npm install -g mographify
Or run directly without installing:
npx mographify record path/to/file.html
Usage
Record an HTML mograph → MP4
mographify record <file.html>
mographify record <file.html> --out ~/Movies
Screenshot every scene for QA review
mographify qa <file.html>
Screenshots land in /tmp/mographify-qa/scene-N.png. Open them to visually inspect before recording.
Inspect a mograph file
mographify info <file.html>
Full pipeline with Claude Code (content → MP4)
The CLI handles the deterministic parts: QA and recording.
The generative part (content brief → HTML) is handled by a Claude Code skill that uses Claude's AI to map your content to the right visual templates, pick the right pillar color, and generate the HTML.
Install the Claude Code skill:
cp claude-skill/SKILL.md ~/.claude/skills/mographify/SKILL.md
Then in Claude Code:
/mographify
## TOPIC
Your topic here
## HOOK
Your hook variant...
## PREMISE
Your core argument...
## CONTENT + POINTS
**Point 1:** ...
**Point 2:** ...
## WHY IT'S RELEVANT
...
Claude maps each section to a visual template, generates the HTML, runs the QA loop (screenshots → vision review → fix loop up to 3 passes), and records the final MP4.
HTML format
Mographs are self-contained HTML files. They run in any browser. The format:
<!-- Stage: 1080×1920 fixed -->
<main class="stage">
<section class="scene active" data-scene="0">
<div class="scene-label">HOOK</div>
<div class="counter">1/5</div>
<div class="visual">
<svg viewBox="0 0 1000 1300">
<!-- CSS-animated elements with .anim .pop .up .left .right .draw classes -->
<!-- style="--d:300ms" sets the animation delay -->
<!-- data-pulse marks the accent element for idle breathing -->
</svg>
</div>
</section>
</main>
<script>
const DURATION = 10000; // ms per scene
const EXIT_DURATION = 1200; // ms for exit animation
// JS controller handles timing, exit transitions, keyboard nav
window._mographStop = () => clearInterval(timer); // QA hook
</script>
Animation classes:
.anim— fade up entry.pop— scale from 0.72.up— translate from below.left/.right— translate from side.draw— stroke dash revealdata-pulse— marks accent element for idle breathing
Exit: When a scene's timer fires, .exiting class is added. CSS !important rules override entry animations with reverse-direction exits (translate up, scale down, undraw).
Design rules
These are enforced by the QA loop:
- Background:
#000000always - Text/shapes:
#ffffff - Accent: exactly one pillar color per scene
- Fonts:
Helvetica Neue700 for labels,Georgiaitalic for statements - SVG viewBox:
0 0 1000 1300always - Node labels: 1–4 words max
- All elements fully within SVG bounds
- No gradients, no decorative noise
Architecture
content brief
│
▼ (generative — Claude AI)
scene plan + template mapping
│
▼ (generative — Claude AI)
HTML file (/tmp/mographify-*.html)
│
▼ (deterministic — Playwright)
QA screenshots (per-scene PNGs)
│
▼ (deterministic — vision review, up to 3 passes)
HTML fixes
│
▼ (deterministic — Playwright + ffmpeg)
MP4 (~Desktop/slug.mp4)
License
MIT