AgentsKit: The Complete JavaScript Toolkit for Building AI Agents

Introduction

AgentsKit.js is presented as the definitive JavaScript toolkit for building AI agents. It offers a lightweight 10KB core, a modular architecture with 22 focused packages, and zero lock-in, ensuring flexibility and reusability. Designed to address the challenges of integrating disparate libraries, AgentsKit provides a cohesive ecosystem for developing everything from basic chat UIs to full-fledged autonomous agents.

Why Use AgentsKit and Key Benefits

Building AI agents in JavaScript often involves combining multiple incompatible libraries, leading to complexity and reusability issues. AgentsKit solves this by offering a small, contracted, and composable solution. Its core benefits include a minimal 10KB gzipped core with zero runtime dependencies, first-class support for agent runtimes (ReAct, tools, skills, delegation, memory, RAG), and easy provider swapping. The ecosystem extends beyond just the package, featuring a Registry for ready-to-use agents and a Playbook for best practices, enabling developers to build, evaluate, and operate agents efficiently. AgentsKit also provides UI surfaces for React and Terminal, making it versatile for various applications.

Installation

To get started with AgentsKit, you can install the core and runtime packages using npm:

npm install @agentskit/core @agentskit/runtime tsx

Examples

Here's a quick start example to create your first agent without requiring an API key:

import type { AdapterFactory } from '@agentskit/core'
import { createRuntime } from '@agentskit/runtime'

const localAdapter: AdapterFactory = {
  createSource(request) {
    const task = request.messages.at(-1)?.content ?? 'your task'

    return {
      async *stream() {
        yield {
          type: 'text' as const,
          content: `Agent ready. I received: ${task}`,
        }
        yield { type: 'done' as const }
      },
      abort() {},
    }
  },
}

async function main() {
  const runtime = createRuntime({ adapter: localAdapter })
  const result = await runtime.run('Plan my first production agent')
  console.log(result.content)
}

void main()

Run it with:

npx tsx agent.ts

For a more advanced autonomous agent with tools, AgentsKit simplifies the process significantly:

Before (typical "JS agent" stack):

// Pick your favorite: LangChain, raw fetch, Vercel AI SDK + custom runtime,
// MCP client + custom UI, manual ReAct loop, hand-rolled streaming...
// Then wire memory. Then wire tools. Then wire delegation. Then debug.

After (AgentsKit):

import { createRuntime } from '@agentskit/runtime'
import { openai } from '@agentskit/adapters'
import { webSearch, filesystem } from '@agentskit/tools'

const runtime = createRuntime({
  adapter: openai({ apiKey: KEY, model: 'gpt-4o' }),
  tools: [webSearch(), ...filesystem({ basePath: './workspace' })],
})

const result = await runtime.run('Research the top 3 AI frameworks and save a summary')

Links

• GitHub Repository
• Official Documentation
• Discord Community
• Public Roadmap