@pylinka/core

Runtime of Pylinka — a GPU-driven, node-based particle system for PixiJS, built for slot games on real phones.

Design an effect in the node editor (or start from one of 60+ recipes), export the project JSON, and play it back in your game:

import { createCompiledParticles } from '@pylinka/core/gpu';

const fx = await createCompiledParticles(canvas, project); // WebGPU compute, else WebGL2 TF
fx.setEmitter(x, y);
app.ticker.add(() => fx.update(1 / 60)); // or your own rAF loop
fx.setKnob('windPower', 40); // live uniform — no recompile

Or run it inside a pixi.js v8 stage — the backend follows the host renderer (a WebGPU renderer shares its device; a WebGL renderer shares its GL context):

import { registerPylinka, createPylinka } from '@pylinka/core/pixi';

registerPylinka();
const app = new Application();
await app.init({ preference: 'webgpu' });
const fx = await createPylinka(project, { renderer: app.renderer });
app.stage.addChild(fx.systems['sparks'].view);
app.ticker.add((t) => fx.update(t.deltaMS / 1000));

• @pylinka/core/gpu — createCompiledParticles: the graph as generated GPU code, best available backend ('auto' | 'webgpu' | 'webgl2').
• @pylinka/core/webgpu — WebGPU compute backend: emit/update kernels + instanced draw.
• @pylinka/core/webgl2 — compiled WebGL2 transform-feedback backend (fused step shader).
• @pylinka/core/pixi — pixi v8 integration: createPylinka, render pipe, follow(), scene interleaving (peer dependency, optional).
• @pylinka/core/webgl — interpreted WebGL2 engine with editor extras: textured atlas sequences, emission masks, sub-emitters, spline trajectories.
• @pylinka/core — CPU scheduler, knob bus, and timing shared by all backends.

npm i @pylinka/core

Docs, node editor, and the recipe gallery: https://pylinka.schmooky.dev

MIT © pylinka contributors