s-m-r-t

@happyvertical/browser-ai

Client-side AI inference with WebGPU acceleration, ONNX Runtime, and offline-first capabilities.

v0.19.0Browser AIONNXWebGPU

Overview

browser-ai enables client-side AI inference in the browser using ONNX Runtime Web with WebGPU/WebAssembly. Provides embedding generation, text generation, and model management.

Installation

bash 
npm install @happyvertical/browser-ai

Quick Start

typescript 
import { BrowserAI } from '@happyvertical/browser-ai';

// Initialize (loads ONNX Runtime in browser)
const ai = await BrowserAI.create();

// Generate embeddings
const embedding = await ai.embed('Hello world');
// Returns: Float32Array of 384 dimensions

// Semantic search
const documents = [
  { id: 1, text: 'Paris is the capital of France' },
  { id: 2, text: 'Tokyo is the capital of Japan' },
  { id: 3, text: 'Berlin is the capital of Germany' }
];

const results = await ai.semanticSearch('What is the capital of France?', documents);
// Returns documents sorted by similarity

// Generate embeddings
const embedding = await ai.generateEmbedding('Machine learning is fascinating');
console.log(embedding); // Float32Array[384]

Related Modules