Java has run enterprise backends for two decades — but most AI tooling is Python-first. Spring AI closes that gap, bringing LLM integration, vector stores, RAG, tool calling, and MCP into the familiar Spring Boot world. This course takes you from your first chat call to a production-grade, observable AI system.
📚 This is a 7-page course. Mark each page complete and your progress fills in on the Courses page. Use the Next button at the bottom of each page to move through it.
Who this is for
Java/Spring engineers and architects who want to build real AI features — not toy demos — with the patterns, structure, and safeguards production demands. Basic Spring Boot knowledge is assumed; no prior AI/ML background needed.
What you'll learn
- Getting Started — the Spring AI model,
ChatClient, configuration, and your first calls. - Prompting, Structured Output & Tool Calling — reliable outputs and letting the model call your Java code.
- RAG & Vector Stores — ground answers in your own data with embeddings and retrieval.
- MCP & Agents — connect tools via the Model Context Protocol and build agentic flows.
- Choosing & Integrating LLMs — which model for which job, and how to integrate any provider.
- Production Architecture — observability, caching, resilience, evals, and the reference architecture.
The big picture
Spring AI's core idea: a provider-agnostic abstraction (ChatClient, EmbeddingModel, VectorStore) so your code stays the same whether you call Claude, OpenAI, Azure, or a local model. You write once; you swap providers via configuration.
How to use this course
Work through the pages in order — each builds on the last. Code examples use Spring Boot 3 + Spring AI. Pair it with the official Spring AI docs and a scratch project you can run as you go. When you finish, you'll have the mental model and patterns to ship enterprise AI on the JVM.
Ready? Start with Getting Started →