Spring Boot meets AI

Preface: the part you’ll skip but shouldn’t

1. Who this is for
2. How the book is built
3. Two ground rules

Part I — Foundations

Chapter 1: Why “Spring Boot + AI,” and why now?

1. What an LLM is, in one breath
2. Why Java developers specifically should care
3. Where Spring AI fits
4. The thing we’re building toward

Chapter 2: Spring Boot basics, recalibrated for AI

1. The 90-second refresher
2. What’s genuinely different about AI apps
3. A configuration sketch we’ll grow into

Chapter 3: AI and LLM concepts for Java developers

1. The mental model: an LLM is a stateless function
2. The five words you need
3. The tasks you’ll actually build
4. How these map to Spring AI

Chapter 4: Getting started with Spring AI

1. What Spring AI is trying to be
2. Bootstrapping with Spring Initializr
3. Configuration
4. “Hello, LLM”

Part II — Core Spring AI patterns

Chapter 5: Chat and completion APIs

1. The message roles, briefly
2. A chat service worth keeping
3. Streaming, finally

Chapter 6: Prompt engineering and templates

1. The anatomy of a good system prompt
2. Don’t build prompts with string concatenation
3. Three patterns that punch above their weight

Chapter 7: Structured outputs and calling into Java code

1. From prose to POJOs
2. Designing DTOs for extraction

Chapter 8: Embeddings and vector stores

1. What an embedding actually is
2. Vector stores: a database that searches by meaning
3. Ingest, store, search

Chapter 9: Retrieval-Augmented Generation (RAG)

1. The RAG loop
2. The three things that make or break a RAG system

Part III — Building real applications and going to production

Chapter 10: Building a complete AI-powered feature

1. The use case, stated like a ticket
2. Clean architecture, AI edition
3. The assembled service
4. The controller

Chapter 11: Multi-provider and model portability

1. Swapping providers is (mostly) a config change
2. Fallback strategies

Chapter 12: Observability and monitoring for AI calls

1. What you need to see
2. Turn on Actuator and metrics
3. Logging prompts and responses — carefully
4. A practical logging interceptor

Chapter 13: Testing AI-powered services

1. Layer 1: test your integration code, not the model (mock it)
2. Layer 2: contract tests for structured outputs
3. Layer 3: the fuzzy stuff — evaluating actual model behavior

Chapter 14: Security, safety, and responsible use

1. Secrets: the unglamorous foundation
2. Rate limiting and input validation
3. Prompt injection: the new SQL injection
4. Moderation and content filters

Chapter 15: Deploying Spring Boot AI apps

1. Packaging with Docker
2. Resource limits and timeouts
3. Outbound calls: proxies, retries, circuit breakers
4. Where to deploy

Part IV — Advanced topics and future directions

Chapter 16: Agents and tool-using workflows

1. Agent vs. plain tool calling
2. Building tools for an agent
3. Modular RAG, while we’re here

Chapter 17: Complex data pipelines and document ingestion

1. The ingestion pipeline: read, split, embed, store
2. Scheduling ingestion
3. The hard part: updates, deletes, and re-embedding

Chapter 18: Patterns, pitfalls, and “smells” in AI apps

1. Smell 1: Over-reliance on brittle prompts
2. Smell 2: Heavy coupling to one model or provider
3. Smell 3: Latency, cost, and fallback blindness
4. Smell 4: Trusting the model’s output unconditionally
5. Smell 5: Reaching for complexity you don’t need
6. A production-readiness checklist

Chapter 19: Where to go next

1. The map you now hold
2. Where to keep learning
3. Extending the patterns to your own domains
4. Your move
5. Appendix: a one-page cheat sheet