Google ADK for TypeScript

Build Sophisticated
Multi-Agent AI Systems

Enterprise-grade framework for hierarchical agents, tool integration, memory management, and real-time streaming

multi-agent-system.ts
const workflow = await AgentBuilder
  .asSequential([researchAgent, analysisAgent])
  .withTools([GoogleSearch, DataProcessor])
  .withMemory(vectorMemoryService)
  .build();

const result = await workflow.ask(
  "Analyze market trends in AI"
);
Multi-Agent
Tool Integration
Memory Services
Real-time Streaming

Why Choose ADK TypeScript?

Everything you need to build production-ready AI agents with TypeScript's type safety and modern tooling.

AgentBuilder API

Fluent interface for rapid agent creation with zero boilerplate. Create agents in one line or build complex workflows.

Multi-LLM Support

Seamlessly switch between OpenAI, Google Gemini, Anthropic Claude, and more with unified interface.

Production Ready

Built-in session management, memory services, streaming, and artifact handling for enterprise deployment.

Advanced Tooling

Custom tools, function integration, Google Cloud tools, MCP support, and automatic schema generation.

Multi-Agent Workflows

Orchestrate complex workflows with parallel, sequential, and hierarchical agent architectures.

Developer Experience

Excellent DX with TypeScript IntelliSense, comprehensive examples, and intuitive APIs.

Build Agents in One Line

AgentBuilder's fluent interface eliminates boilerplate and lets you focus on building intelligent agents.

Simple yet Powerful

From one-line agents to complex multi-agent workflows, AgentBuilder scales with your needs.

1
Install @iqai/adk package
2
Choose your LLM provider
3
Start building agents

Supported Models: GPT-4, Claude-3.5-Sonnet, Gemini-2.5-Flash, and more

import { AgentBuilder } from '@iqai/adk';

// One-line agent creation
const response = await AgentBuilder
  .withModel("gemini-2.5-flash")
  .ask("What is the primary function of an AI agent?");

// Agent with session and tools
const { agent, runner, session } = await AgentBuilder
  .create("my_assistant")
  .withModel("gpt-4")
  .withDescription("A helpful AI assistant")
  .withInstruction("Provide concise responses.")
  .withTools(new GoogleSearch(), new HttpRequestTool())
  .withQuickSession("my-app", "user-123")
  .build();

// Multi-agent workflow
const workflow = await AgentBuilder
  .create("research_workflow")
  .asSequential([researchAgent, summaryAgent])
  .build();

Ready to Build Your First Agent?

Join developers building the future of AI with TypeScript. Get started with our comprehensive documentation and examples.