Building Live Voice Agents

Chapter 9. Implementing A2A-Compliant Agents

A Note for Early Release Readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be Chapter 9 of the final book. Please note that the GitHub repo will be made active later on.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at [e-mail].

The previous chapter treated Agent-to-Agent (A2A) as a protocol. This chapter turns that protocol into runnable service boundaries: an agent publishes a public card, receives a protocol message, translates that message into its own implementation, and returns a response that the caller can use without knowing the agent’s framework.

To show how these boundaries work in practice, this chapter develops three example implementations. In each, callers depend on agent cards and A2A messages, artifacts, and task status rather than framework-specific internals. Each implementation answers a different integration question:

• An ADK initiator calls a LangGraph greeter through its public A2A surface. Here you see a cross-framework handshake between a LangGraph greeter and a Google ADK initiator.

• A math client sends a question to an ADK-backed specialist without reaching into its tool code. In this example, you find an A2A wrapper around an existing ADK math agent. ...