google/adk-js

What it solves

It provides a code-first framework for developers to build, evaluate, and deploy AI agents. It removes the friction of orchestrating complex agent workflows—ranging from simple tasks to multi-agent systems—by allowing logic to be defined directly in TypeScript for better debugging and versioning.

How it works

ADK allows developers to define agent behaviors, instructions, and tool usage using TypeScript. It leverages Zod for schema validation and type safety. Agents can be composed into various workflow patterns, including sequential, parallel, loop, and routed structures, and can communicate with remote agents via the A2A protocol. It supports running in both Node.js and browser environments.

Who it’s for

TypeScript developers building AI agents for server-side (Node.js) or client-side (browser) applications who want full type safety and programmatic control over agent orchestration.

Highlights

• Multi-Runtime Support: Compatible with ESM, CommonJS, and web bundles for deployment in Node.js or the browser.
• Integrated Tooling: Built-in support for Google Search, Google Maps, Vertex AI Search, and MCP servers.
• Developer Experience: Includes a CLI for scaffolding and deployment, and an interactive web UI for testing and debugging.
• Flexible Orchestration: Supports complex multi-agent delegation and various workflow patterns.