Building a Star Trek Computer with Java 25 and Spring AI

Can you talk to your computer in plain English and have it diagnose itself — Star Trek style? In this session, we’ll build exactly that using the modern Java ecosystem. Starting with OSQuery — a utility that exposes operating system data as SQL tables — we’ll use Spring AI 2 (on Spring Boot 4) and the Model Context Protocol (MCP) to create a natural language interface to your machine. The LLM translates spoken commands like “run a level one diagnostic” into the right SQL queries, executes them through an MCP server written in Java, and streams back a human-readable report. The session covers writing MCP tool methods with Spring AI’s @Tool annotation, building a JavaFX voice interface that uses Whisper for speech-to-text, running concurrent system queries with virtual threads and CompletableFuture, and streaming results back using reactive Flux. We’ll also look at how agentic CLI skills (like those in Claude Code) offer a simpler alternative when a full MCP server isn’t needed. All code is open source and available on GitHub.