Unleashing Creativity with Spring AI
summary
Built a live escape-room chatbot called Overmind using Spring AI plus OpenAI for embeddings and GPT for responses, storing vectors in PGVector and using text-embedding-ada-002; game enforces coins and levels, RAG returns only the top three documents, and a custom level-up advisor grants more knowledge when player queries match curated steps. Initial on‑prem Llama fails due to token size so the system switches to OpenAI; cost falls from about $5/day to ~$1/day through prompt and retrieval pruning; no one reaches level three in the demo.
AI-written summary · about
chapters (14)
- 0:00Overmind and team intro
- 1:06Overmind’s dystopian premise
- 3:11Gameplay mechanics overview
- 5:02AI explosion and business models
- 8:35Private data integration techniques
- 10:20Embeddings and vector databases
- 12:35Spring AI introduction
- 15:49Spring AI pros and cons
- 19:43Spring AI configuration deep dive
- 24:00Spring AI advisors and calls
- 27:31Function calls for business logic
- 28:59RAG pipeline and ingestion
- 30:05Overmind next steps
- 32:16Q&A
related talks
01h 00mSpring AI
02h 43mSpring into the AI Era: A Practical Exploration for Developers (Workshop)
44mReal-World AI Patterns with Spring AI and Vaadin
51mIntelligent Applications with Spring AI
52mIntelligent Beans with Spring AI
55mSpring into the AI Era: A Practical Exploration for Developers
49mBuilding AI Agents with Spring & MCP
54mAI 202: Next-Level AI Mastery for Java Developers