Mayelana ne-PinePaper Studio

Ukwakha ikusasa lemidwebo enomdlandla: esekwe esiphequluli, ye-AI yemvelo, netholakala kubo bonke.

Umgomo Wethu

I-PinePaper Studio iletha intando yeningi ekuklameni kwemidwebo enomdlandla.

Okusenza Sihluke

🤖

Umklamo we-AI Wemvelo

Lwakhiwe kusukela phansi ukusebenza ne-AI.

🎬

Khipha MP4 & GIF

Khipha ividiyo ye-MP4 ye-Instagram, TikTok, YouTube. I-GIF yokuthumela imilayezo.

⏱️

Timeline & Keyframes

Professional timeline editor with keyframe-based animations. Create choreographed sequences with precise timing control, 5 easing functions, and color interpolation.

🌍

Komhlaba Wonke Ngokuzenzakalelayo

Usekelo lwezilimi ezingu-59 nazo zonke izinhlelo zokubhala. Dala okuqukethwe kunoma yiziphi izethameli.

🚀

Inkululeko Yesiphequluli

Akudingeki ukufakwa. Isebenza kunoma iyiphi idivayisi enesiphequluli sesimanje.

📋

Umtapo Wamathempulethi

Qala ukudala ngamathempulethi aselungisiwe. Enzelwe imithombo yezindaba, ibhizinisi, nemfundo.

🎨

Isixhumi Esikabili

Intuitive UI for human designers and a powerful API for AI agents and developers. Two interfaces, one mission: empowering creativity.

🖨️

Umklamo Olungele Ukushicilela

Create professional print materials with A3, A4, A5, Letter, Legal, and Tabloid presets at 300 DPI. Export to PDF with bleed and trim marks, or high-DPI PNG (up to 600 DPI) for commercial printing.

v0.4
🦴

Ukwakha Isimo

Umfanekiso wamathambo namathambo, ukusabalalisa kwe-FK, ne-FABRIK IK.

v0.4
🎭

Ukuxuba Nokuhlangana

Per-item and per-group blend modes with 7 presets (ghost, neon, shadow, glow, xray, dreamy, vintage). Animated transitions, interactive hover/click modes, and proximity-based blend reactions.

v0.4
✂️

Izindlela Zokusika Nesifonyo

Animate stroke drawing with trim paths for "line draw" effects. Advanced masking with clip/subtract/intersect modes, mask stacking, animated reveals (14 presets), and vertex animation.

v0.4
🔗

Ubudlelwano Bokuqhubeka

New relation types: driven_by (property linking), wiggle (procedural noise), and time_expression (math expressions with t). Plus nested compositions (precomps) with independent local timelines.

v0.4
🌐

Indwangu Engenamkhawulo

Draw anywhere without boundaries. Unbounded canvas mode removes all size restrictions — pan and zoom freely across an infinite workspace.

v0.4
🖌️

Amathuluzi Okudweba

6 professional drawing tools: pen, ink brush, marker, spray can, eraser, and fill bucket. Each with custom cursors and optimized stroke algorithms.

v0.4
🧊

Ukuboniswa Kwemibono

Bonisa izinhlobo (ikhubu, ibhola, isilinda, ithori, ikoni) kusithombe ngezinhlobo ezingu-5 zokubonisa.

v0.5
🎛

Izihluzi Zezithombe ze-GPU

Real-time image processing powered by WebGPU compute shaders with WebGL2 fallback. HSL adjustment, brightness, contrast, posterize, edge detection, color tint, vignette, blur, and filter chaining in a single GPU pass.

v0.5
✂

Ukuhlukanisa Isithombe

Freehand lasso tool to select and extract image regions at full native resolution. 8 cutout style presets (papercut, postcard, sticker, poster, mono, sketch, warm, cool) combine GPU filters with shadow effects.

v0.5
📐

Injini Yezibalo Nesayensi

Numerical computing powered by math.js — ODE solvers (RK4/RK45), FFT spectrum analysis, function plotting, parametric curves, dynamic system simulation, and parametric surfaces in perspective. Real math, animated on canvas.

Compliance & Trust

Level AA conformance, W3C WAI Web Content Accessibility Guidelines 2.1

WCAG 2.1 AA

Accessibility shadow tree for screen reader canvas navigation. Keyboard navigation, live region announcements, skip links, contrast auditing, and motion sensitivity checks. Uses pp:PinePaper ontology for semantic item descriptions.

🔒

Security Headers

HSTS with preload, Content Security Policy, X-Frame-Options, X-Content-Type-Options, Permissions-Policy. No camera, microphone, geolocation, or payment API access. Frame-ancestors restricted.

👁

Privacy by Design

100% client-side processing. No server uploads, no cookies, no user data collection. All designs stay in your browser. Export locally — nothing leaves your device unless you choose to share.

Ubuchwepheshe Namandla

Ubuchwepheshe Obuyinhloko

  • ▹ Paper.js: Vector graphics manipulation and rendering
  • ▹ HTML5 Canvas: Real-time animation and export
  • ▹ Vanilla JavaScript: Lightweight, fast, no dependencies
  • ▹ Timeline Engine: Keyframe interpolation with easing functions
  • ▹ Rigging Engine: Skeletal animation with FK/IK solvers and pose interpolation v0.4
  • ▹ Design Knowledge Graph: Semantic ontology with 25+ relation types, 35+ math functions, and type hierarchy v0.4
  • ▹ Rust/WASM Runtime: Optional WebAssembly acceleration for widget computation — 10-100x faster relation processing, no GC pauses. Falls back to JS when WASM isn't available v0.5
  • ▹ pp:PinePaper Ontology: 55-token semantic vocabulary (items, relations, animations, masks, generators) — the language of PinePaper animations. Embedded in every exported widget v0.5
  • ▹ Sentry: Error tracking and performance monitoring

Izinhlobo Zokukhipha

  • ▹ WebM Video: VP9 codec - best quality for animations
  • ▹ MP4 Video: H.264 via WebCodecs API - ideal for social media
  • ▹ GIF Animation: Via gif.js Web Workers - perfect for messaging
  • ▹ PDF for Print: 300 DPI with bleed (3-5mm) and trim marks - professional printing
  • ▹ High-DPI PNG: Up to 600 DPI for commercial print quality
  • ▹ Animated SVG: Native SVG animations with SMIL
  • ▹ Interactive Widget: Self-contained HTML with scene-specific tree-shaken code (~8-20KB). WASM-accelerated when available. Embed anywhere — no CDN, no API, no dependencies v0.5
  • ▹ Lottie JSON: Industry-standard animation format. Auto-bakes procedural animations into native keyframes for any Lottie player v0.4

Izinhlobo Zomdlandla

Pulse Rotate Bounce Fade Wobble Swing Shake Jelly Umshini Wokubhala Custom Paths Keyframe Animation ✨ Trim Paths Spatial Motion Paths Path Morphing 13 Deform Presets 100+ Animation Presets

Isistimu Yemiphumela

10 Visual Effects Sparkle Blast Fire Smoke Rain Snow Confetti Ripple Glow Electric 14 Mask Presets 19 Blend Presets 12 Blend Modes

On-Device AI EXPERIMENTAL

PinePaper supports client-side LLMs that run entirely in your browser — no account, nothing sent to a server. Chrome's built-in Gemini Nano, or any other in-browser model, can follow this experimental process to let its agent talk to PinePaper's tools directly.

The model emits PinePaper tool calls constrained by a grammar (a JSON schema on Chrome's Prompt API, an EBNF grammar on WebLLM), so the output is always valid and runs straight onto the canvas — a prompt becomes a scene in one step. Find it under AI / Code → Assistant in the editor. Being on-device, it's free and private; quality tracks whatever small model your browser provides, while the Cloud tab and MCP below give you larger models for higher fidelity.

PinePaper MCP Server

Use PinePaper with Claude Desktop and other AI assistants via the @pinepaper.studio/mcp-server npm package. Create animated graphics using natural language:

Pair MCP with the official Claude Skill

The Skill teaches Claude when to reach for PinePaper and how to drive it correctly. Install once and ask in plain English — "animate this logo", "make a pulsing LIVE badge" — and Claude routes through PinePaper instead of generating HTML/CSS.

Skill docs →

Quick Setup

1. Install globally:

npm install -g @pinepaper.studio/mcp-server

2. Add to Claude Desktop config:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json 

{
  "mcpServers": {
    "pinepaper": {
      "command": "npx",
      "args": ["-y", "@pinepaper.studio/mcp-server"]
    }
  }
}

3. Try it in Claude Desktop:

"Create a red pulsing text that says HELLO" "Make earth orbit the sun"
Iphakeji ye-npm Nika Inkanyezi ku-GitHub Amadokhumenti Aphelele →

Available Tools

  • • Create text, shapes, graphics
  • • Add relations (25+ types: orbits, follows, wiggle, etc.)
  • • Keyframe & preset animations
  • • Character rigging (skeletons, bones, IK/FK)
  • • Blending presets & transitions
  • • Diagrams, maps, font creation
  • • Generate procedural backgrounds (14 generators)
  • • Vertex deformation (13 presets: fold, twist, wave, etc.)
  • • Export SVG, training data

Supported AI Platforms

  • • Claude Desktop
  • • Any MCP-compatible client
  • • Custom integrations via API

Igrafu Yolwazi Lomklamo

Unveiled in v0.4, the Design Knowledge Graph is a semantic reasoning layer that maps relationships between visual elements. Unlike traditional tools that only store pixel or vector data, PinePaper understands the intent behind the design—whether an object "orbits" another, "points at" a target, or "follows" a path.

Le ndlela entsha ivumela abameli be-AI ukucabanga ngomdlandla oyinkimbinkimbi.

Attributed Research

The Design Knowledge Graph architecture draws on ideas from these peer-reviewed papers:

For Developers & AI Agents

PinePaper exposes a comprehensive global API (window.PinePaper) that enables programmatic control of the entire application. AI assistants and developers can:

  • • Create and manipulate text, shapes, and imported SVG elements
  • • Apply animations (9 presets) and effects (sparkle, blast)
  • • Create keyframe-based animations with timeline control
  • • Export to MP4, GIF, SVG, or PNG via API
  • • Use declarative relations (orbits, follows, attached_to)
  • • Rig characters with skeletal animation and pose keyframes v0.4
  • • Create nested compositions (precomps) with local timelines v0.4
  • • Build automated workflows and batch processing systems

Example API Usage:

// Create animated text
const text = app.create('text', {
  content: 'Hello World', x: 400, y: 300, fontSize: 48, color: '#4a9eff'
});

// Apply animation preset
app.animate(text, { animationType: 'pulse' });

// Keyframe animation with easing
app.create('circle', {
  x: 400, y: 300, radius: 50, color: '#ef4444',
  animationType: 'keyframe',
  keyframes: [
    { time: 0, properties: { y: 100, opacity: 0 } },
    { time: 1, properties: { y: 300, opacity: 1 }, easing: 'bounce' }
  ]
});

// Declarative relations
app.addRelation(moonId, earthId, 'orbits', { radius: 100, speed: 0.5 });

Umbono Wethu

Sakha inkundla yokuqala yemidwebo enomdlandla ye-AI yemvelo.

Umgomo wethu ukwenza imidwebo enomdlandla ibe lula njengokubhala umbhalo.

✨ Yini Elandelayo?

v0.5 shipped data-driven charts, vision-puppeteering (webcam → rigged character), the measurement system with rulers + snap, glTF / BVH export, and the Claude Skill + MCP plugin marketplace. Up next: deeper agent integrations, animation state machines as a first-class primitive, and a growing template library across art, education, and motion-design domains.

Buka Imephu & Yabelana Bika Iphutha

Community & Open Development

PinePaper Studio is developed transparently with our community. We welcome feedback, feature requests, and bug reports. Our development is guided by real user needs and the evolving landscape of AI-assisted creativity.

💬 Yabelana Ngempendulo 🗺️ Buka Imephu

Yamahhala futhi Evulekile

PinePaper Studio is Freeware v1.0 (currently v0.5.3-beta) and free to use for personal and commercial purposes. Create unlimited animations, export unlimited files, and use them in any project without restrictions. We believe powerful creative tools should be accessible to everyone.

59
Izilimi Ezisekelwayo
25+
Izinhlobo Zobudlelwano
111+
API Methods
7
Izinhlobo Zokukhipha

Licensing & Open Source Credits

PinePaper Studio is built on the shoulders of giants. We are grateful to the maintainers of these open-source projects, without which PinePaper would not exist. Each library is bundled under its original license — please refer to the linked source for full terms.

PinePaper Studio itself

  • ▹
    Application code Proprietary Freeware
    Free to use; source code is not licensed for redistribution, modification, or forking. See LICENSE and Terms of Service.
  • ▹
    PinePaper Ontology (pp: vocabulary) CC0 1.0 Public Domain
    The semantic vocabulary and Knowledge Graph are dedicated to the public domain. Use pp: triples in your own canvases, exports, training data, or models with no attribution required. See /ontology/.
  • ▹
    MCP Server MIT
    The Model Context Protocol server is open source and may be forked, modified, and redistributed. See github.com/pinepaper/mcp-server.
  • ▹
    PineWidget runtime Embeddable
    The standalone widget player ships unobfuscated and is intended for third-party embedding.
▹
Paper.js MIT
Vector graphics rendering foundation
▹
math.js Apache-2.0
Expression evaluation, ODE solvers, numerical analysis
▹
opentype.js MIT
Font glyph parsing for the Font Studio
▹
planck-js MIT
Box2D physics engine for canvas item simulation
▹
d3-geo ISC
Geographic projections for the Map System
▹
d3-geo-projection ISC
Extended map projections (Robinson, Mollweide, Winkel)
▹
topojson-client ISC
TopoJSON-to-GeoJSON converter for map data
▹
gif.js MIT
Animated GIF export via Web Workers
▹
jsPDF MIT
PDF export with bleed and trim marks
▹
@sentry/browser MIT
Error tracking (production build only)
▹
DOMPurify Apache-2.0 OR MPL-2.0
Cure53's XSS sanitizer — runs on every imported SVG before Paper.js sees it, so a malicious file can't smuggle scripts or event handlers into the editor.
▹
@observablehq/plot ISC
Data visualization scales and transforms for the Chart System
▹
@huggingface/transformers Apache-2.0
Vendored browser ESM build + ONNX runtime WebAssembly. Powers PinePaper's on-device object detection (the "Detect objects" / open-vocabulary "Find" features) — inference runs on your GPU (WebGPU) or CPU; the image never leaves your device.
▹
DETR ResNet-50 Apache-2.0
Meta AI's DETR object-detection model (80 COCO classes). Runs in the browser via transformers.js (fp16 on WebGPU, q8 on CPU) behind the "Detect objects" button — finds objects in an image so you can label, relate, or composite onto them.
▹
OWL-ViT (base-patch32) Apache-2.0
Google Research's open-vocabulary detector — finds anything you describe in text, not just fixed classes. Powers the "Find" feature's AI fallback (zero-shot detection inside a photo). Large model (~150 MB), so it's strictly opt-in and downloads once on first use; runs fully on-device.
▹
Qwen2.5-1.5B-Instruct Apache-2.0
Alibaba's Qwen2.5 instruction-tuned LLM. Quantized (q4f16) and run fully on-device via WebGPU to power the PinePaper client AI assistant — prompts never leave your device. A smaller Qwen2.5-0.5B-Instruct (also Apache-2.0) is used as the low-storage fallback.
▹
WebLLM (MLC) Apache-2.0
MLC-AI's in-browser LLM runtime — loads the Qwen2.5 weights and runs inference on WebGPU for the on-device assistant.
▹
LibreTranslate AGPL-3.0 Build-time
Self-hosted machine-translation API used at build time to translate the editor UI, User Guide, and Knowledge Hub into the supported languages. Not bundled or network-served by the app — only its output (the translated text) ships.
▹
Argos Translate MIT Build-time
The open neural translation engine (and open language models) that LibreTranslate runs under the hood for our build-time localization pipeline.

Each linked project's repository carries its own license file. We ship obfuscated production builds so a local node_modules tree isn't in the distribution; we keep this list above and these upstream links current instead. If we have missed attributing a project, please let us know.