A Beginner’s Guide to Spring AI Application Development
A Juejin article presents a beginner-oriented introduction to AI application development for backend developers familiar with Spring Boot. It explains architecture, RAG, Tool Calling, MCP, Spring AI Advisors, memory, vector retrieval and Agent/ReAct concepts, alongside runnable code and five implementation pitfalls. The article also compares Alibaba Cloud Bailian and DashScope and discusses how models can be accessed through cloud services, self-hosted deployments, platforms, clients and code.
This guide organizes several commonly confused AI application-development concepts into a Spring AI-oriented learning path for backend developers.
适合人群：会 Spring Boot、想上手 AI 应用开发，但被一堆新名词劝退的后端同学。 读完你会得到：一张可背的架构图、一份概念对照表、一段能跑的最小代码、以及 5 个我真实踩过的坑。
一、先做个自测：这 5 个问题你能答上来几个？
不用查资料，凭直觉答：
- 大模型接入的「2 种途径、3 种方式、4 种调用」，分别指什么？
- 「百炼」和「灵积 DashScope」是什么关系？写代码时该用哪个？
- RAG 的 ETL 和 检索，哪个是离线跑的、哪个是每次请求都跑的？
- Tool Calling 和 MCP 有什么区别？为什么有了前者还要有后者？
- 模型调用工具时，为什么日志里会出现两轮 Request/Response？
如果 5 题里有 2 题答不上来，这篇就是给你写的。
#Spring AI#Spring Boot#Alibaba Cloud Bailian#DashScope#Qwen#DeepSeek#Ollama#PGvector#LangChain4j