Agentary JS

Run quantized small language models directly in the browser with WebGPU and WebAssembly, featuring built-in support for agentic workflows.

Features

🚀 Browser-Native - Run LLMs directly in the browser without server dependencies
⚡ WebGPU Acceleration - Leverage WebGPU for high-performance inference
🤖 Agentic Workflows - Create and execute complex multi-step agent workflows
🧠 Memory Management - Smart context compression and pruning for long conversations
🛠️ Function Calling - Built-in support for tool/function calling
📊 Multi-Model Support - Use different models for chat, tool use, and reasoning
📡 Lifecycle Events - Built-in event system for monitoring and debugging

Quick Start

Install via npm:

npm install agentary-js

Basic example:

import { createSession } from 'agentary-js';
 
const session = await createSession({
  models: {
    chat: {
      name: 'onnx-community/gemma-3-270m-it-ONNX',
      quantization: 'q4'
    }
  },
  engine: 'webgpu'
});
 
for await (const chunk of session.createResponse({
  messages: [{ role: 'user', content: 'Hello!' }]
})) {
  process.stdout.write(chunk.token);
}
 
await session.dispose();

Why Agentary JS?

Browser-Native AI

Run powerful language models directly in users’ browsers - no server costs, maximum privacy, and instant responses.

Agentic Workflows

Build sophisticated multi-step AI agents that can think, plan, use tools, and adapt - all running client-side.

Production Ready

Built with observability, error handling, memory management, and performance optimization from the ground up.

Browser Support

WebGPU: Chrome 113+, Edge 113+, Firefox with WebGPU enabled
WebAssembly: All modern browsers
Minimum Requirements: 4GB RAM recommended for small models

Get Started

Ready to build? Check out the Getting Started guide.

Community

Installation