Building AI Agents Using Google Agent Development Kit (ADK)

Building AI Agents Using Google Agent Development Kit (ADK)

The Google Agent Development Kit (ADK) is an open-source framework enabling the creation of AI agents capable of orchestrating complex workflows between Large Language Models (LLMs) and external tools within the Java ecosystem. This tutorial demonstrates building a basic AI agent named Baelgent with a unique persona, conversation history, and a custom tool for data retrieval.

Why This Matters

Current LLM applications often fall short of real-world needs, requiring complex orchestration beyond simple question-answering. Without frameworks like ADK, developers face significant challenges in managing state, integrating tools, and ensuring reliable interactions, leading to increased development time and potential failure rates in production systems.

Key Insights

• Gemini API key required: Running the application necessitates passing a Gemini API key via the GOOGLE_API_KEY environment variable.
• Function calling: The ADK supports function calling, enabling LLMs to invoke external Java methods during conversations.
• Temporal integration: Temporal is used by companies like Stripe and Coinbase for reliable workflow orchestration, similar in concept to the ADK’s agent capabilities.

Working Example

@Service
class AgentService {
private final InMemoryRunner runner;
private final ConcurrentMap<String, Session> inMemorySessionCache = new ConcurrentHashMap<>();
AgentService(BaseAgent baseAgent) {
this.runner = new InMemoryRunner(baseAgent);
}
UserResponse interact(UserRequest request) {
UUID userId = request.userId() != null ? request.userId() : UUID.randomUUID();
UUID sessionId = request.sessionId() != null ? request.sessionId() : UUID.randomUUID();
String cacheKey = userId + ":" + sessionId;
Session session = inMemorySessionCache.computeIfAbsent(cacheKey, key ->
runner.sessionService()
.createSession(runner.appName(), userId.toString(), null, sessionId.toString())
.blockingGet()
);
Content userMessage = Content.fromParts(Part.fromText(request.question()));
StringBuilder answerBuilder = new StringBuilder();
runner.runAsync(userId.toString(), session.id(), userMessage)
.blockingForEach(event -> {
String content = event.stringifyContent();
if (content != null && !content.isBlank()) {
answerBuilder.append(content);
}
});
return new UserResponse(userId, sessionId, answerBuilder.toString());
}
}

Practical Applications

• Customer Support Chatbots: Companies can use ADK to build intelligent chatbots that handle complex customer inquiries and integrate with backend systems.
• Pitfall: Overly complex prompts can lead to unpredictable agent behavior and increased costs due to excessive LLM token usage.

References:

• https://www.baeldung.com/java-google-agent-development-kit