Bootiful Spring AI
summary
Builds a dog-adoption assistant that actually runs: a Spring Boot app using Bedrock (Nova models) for chat and embeddings, PGVector in Postgres as the vector store, and Spring AI’s RAG, memory advisor and skills-jars for Pooch Palace facts. Tool calling moves scheduling into an external MCP tool calling server authenticated via OAuth, startup creates embeddings, Grafana shows token and latency metrics, and the demo has live hiccups but completes successfully, proving production-ready integration.
AI-written summary · about
chapters (16)
- 0:00Welcome to Spring AI
- 2:19MCP server for the book
- 4:05Dog adoption AI assistant
- 13:25System prompts for Pooch Palace
- 15:14AI skills for domain knowledge
- 17:44Tool calling and agent integrations
- 21:36Chat memory for persistence
- 24:22RAG with vector stores
- 27:13Structured outputs and typed responses
- 28:39Adoption tool and local agent tools
- 30:46MCP servers for distributed AI
- 33:14Securing AI applications
- 38:07OAuth client setup for MCP
- 39:49Testing secure AI tool calls
- 40:10Observability and production
- 41:12Call to action for Java developers
related talks
40mBuilding AI Agents with Spring & MCP
01h 00mSpring AI
49mBuilding AI Agents with Spring & MCP
54mBootiful Spring Boot 3.x
53mHands On Building Agents
55mSpring into the AI Era: A Practical Exploration for Developers
27mBuilding AI Agents with Spring & MCP
45mThe Overwhelmed Spring Developer's Guide To AI