woolball-client 🧶

Turn any open tab into an AI compute node

woolball-client establishes a connection between a user's browser and a running woolball-server instance. It utilizes WebSocket to receive job requests, executes the specified model locally using WebGPU or WASM, and transmits the outcome back to the server in real-time.

⚠️ Important: Before using this library, make sure you have the woolball-server running

AI Tasks

| Provider | Task | Models | Status | |----------|------|--------|--------| | Transformers.js | Speech-to-Text | ONNX Models | ✅ Implemented | | Transformers.js | Text-to-Speech | ONNX Models | ✅ Implemented | | Kokoro.js | Text-to-Speech | ONNX Models | ✅ Implemented | | Transformers.js | Translation | ONNX Models | ✅ Implemented | | Prompt API | Translation | Gemini Nano | 🧪 Experimental | | Transformers.js | Text-Generation | ONNX Models | ✅ Implemented | | WebLLM | Text Generation | MLC Models | ✅ Implemented | | MediaPipe | Text Generation | LiteRT Models | ✅ Implemented | | Prompt API | Text Generation | Gemini Nano | 🧪 Experimental | | Transformers.js | Image-Text-to-Text | ONNX Models | ⚠️ Partial | | Prompt API | Image-Text-to-Text | Gemini Nano | 🧪 Experimental |

Browser API Tasks

| API | Task | Description | Status | |-----|------|-------------|--------| | Canvas API | Character-to-Image | Converts a character to an image | ✅ Implemented | | Canvas API | HTML-to-Image | Converts HTML content to an image | ✅ Implemented |

Quick Start

1. Install the package:

npm install woolball-client

2. Import and use in your code:

import Woolball from 'woolball-client';

// Initialize with a client ID
const woolball = new Woolball('your-client-id', 'ws.server.com'); // ws://localhost:9003 by default

// Listen for task events

Usage Options

Web Application

See the usage directory for a complete React-based web application example that demonstrates how to integrate the Woolball client into a web application.

Chrome Extension

A Chrome extension is available in the chrome-extension directory, allowing users to contribute their browser's computing resources to a Woolball server while browsing.

To build and use the Chrome extension:

# Build the Woolball client library
npm run build:all

# Install extension dependencies
cd chrome-extension
npm install

# Build the extension
npm run build

Then load the extension in Chrome from the chrome-extension/dist directory.

Demo Pages

Simple demo pages are available in the demo directory:

# Serve the demo pages
npm run serve:demo

Development

# Install dependencies
npm install

# Build the library
npm run build:all

# Run tests
npm test

# Run end-to-end tests
npm run test:e2e

License

This project is licensed under the MPL-2.0 License - see the LICENSE file for details.