// The AI will automatically call this Java function when needed. Spring AI auto-instruments all AI calls. Add micrometer-tracing-bridge-otel to your pom.xml , and you get a Grafana dashboard showing token usage, latency, and cost per request. This is critical for "in production" vs. "in action." Part 7: Where to Find the Best "Spring AI in Action" Resources (Sept 2025) Given the rapid evolution, use this matrix to find exactly what you need:
Enter . This new addition to the Spring ecosystem provides an abstraction layer for AI models, similar to how Spring Data abstracts databases.
Spring AI is not a passing trend; it is the future of enterprise Java. The "action" is happening right now, in commits, in PRs, and in those tiny, powerful code snippets that turn a PDF into a smart assistant. Your journey starts with a git clone and a dot (period) to open the PDF. spring ai in action pdf github
aka.ms/spring-ai-starters (Microsoft and VMware collaboration repo) – Often ranks better than Google for practical demos. Conclusion: From PDF to Production The search for "spring ai in action pdf github" reveals a specific developer need: Actionable, executable knowledge. You don't want marketing hype. You want to see the @Service annotation next to an ChatClient , and you want a PDF you can read on the train.
public String ask(String question) // 1. Find relevant PDF chunks List<Document> relevantDocs = vectorStore.similaritySearch(question); // 2. Create the system prompt with context var systemPrompt = """ You are a helpful assistant. Answer using only the provided context. Context: %s """.formatted(relevantDocs.toString()); // 3. The "In Action" call return chatClient.prompt() .system(systemPrompt) .user(question) .call() .content(); // The AI will automatically call this Java
Then, generate the official PDF from the spring-ai-docs module. Keep it on your desktop.
Clone it. Run ./mvnw spring-boot:run . Open localhost:8080 . Ask a question. This is critical for "in production" vs
record Actor(String name, Integer age) {} Actor actor = chatClient.prompt() .user("Generate an actor from the 1990s") .call() .entity(Actor.class); // No JSON parsing boilerplate! From spring-tips repo: