May 11
Loading video player...
The Model Context Protocol (MCP) is the standard integration layer between AI models and your backend services. With Spring AI 1.1+, building an MCP server is as simple as writing a regular Spring service — one dependency, three annotations, zero glue code. In this video you'll build a working product catalogue tool that Claude or GPT can call directly from natural language — end to end in 20 minutes. What you'll learn: ✅ What MCP is and why it replaces custom AI integrations ✅ The three-tier architecture: Host → Client → Server ✅ Building an MCP Server with @McpTool and @McpToolParam annotations ✅ Wiring the MCP Client — YAML config + ToolCallbackProvider + ChatClient ✅ A live demo: asking "What is the price of SKU-001?" and watching the tool fire ✅ 4 production gotchas: SSE deprecation, annotation versions, K8s stateless mode RESOURCES • Spring AI MCP docs: https://docs.spring.io/spring-ai/reference/guides/getting-started-mcp.html • MCP Java SDK: https://github.com/modelcontextprotocol/java-sdk • Streamable HTTP docs: https://docs.spring.io/spring-ai/reference/api/mcp/mcp-streamable-http-server-boot-starter-docs.html • Full source code: https://github.com/brajamony/brajamony/tree/main/mcp-demo-ollama/mcp-demo-ollama