google/adk-go

An open-source, code-first Go toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.

What it solves

It provides a structured, code-first approach to building AI agents, moving away from fragile prompt-based orchestration to a more stable software engineering approach. It simplifies the process of creating, evaluating, and deploying agents that can perform tasks using a variety of tools.

How it works

ADK is a modular framework that allows developers to define agent logic, tools, and orchestration directly in Go. It is model-agnostic and deployment-agnostic, meaning it can work with various LLMs (though optimized for Gemini) and can be deployed in any containerized environment, such as Google Cloud Run.

Who it’s for

Developers who want to build cloud-native AI agent applications using Go, prioritizing performance, concurrency, and software development best practices like testability and versioning.

Highlights

Code-First Design: Logic and orchestration are defined in Go code rather than through configuration files or prompts.
Multi-Agent Systems: Supports the creation of scalable applications by composing multiple specialized agents.
Tool Integration: Offers a rich ecosystem of pre-built tools and the ability to integrate custom functions.
Cloud-Native Ready: Designed for easy containerization and deployment to cloud environments.