Google has broadened its Agent Development Kit (ADK) by introducing native support for the Go programming language, giving developers a new option for creating production-grade AI-driven agent applications.

The update, announced on November 7, brings an idiomatic Go implementation to the ADK ecosystem. According to Google’s developer relations team, the new Go edition emphasizes performance, strong typing, and Go’s built-in concurrency model to help engineers build reliable and scalable agent-based systems. The full implementation is available as open-source code on GitHub.

With the addition of Go, the ADK now offers three language environments—Go, Java, and Python—allowing teams to choose the ecosystem that best fits their infrastructure or expertise.

Highlights of the Go-based ADK release

• Extensive tooling and integrations: The kit includes predefined tools, customizable functions, OpenAPI specifications, and support for connecting to various Google services.
• Code-centric workflow: Developers can express agent behaviors, operational logic, and orchestration patterns directly in code, improving maintainability, test coverage, and version control.
• Modular agent architecture: The framework enables the assembly of multiple specialized agents into scalable, hierarchical systems suited for more complex applications.
• Integrated development interface: A built-in UI assists with inspecting, evaluating, debugging, and demonstrating agent behavior during development.
• Agent2Agent (A2A) protocol compatibility: The ADK supports communication between agents, allowing a main agent to coordinate, delegate tasks, and manage specialized sub-agents.

By shifting orchestration mechanics and tool usage into the code layer, Google aims to give developers more predictable debugging, clearer lifecycle management, and broader deployment flexibility when working with agentic AI applications.