I wanted something faster than gradle build for my coding agent, so I made a Java checker in Rust

Developer Mihai Daniel has introduced 'grounds,' an experimental static analysis tool for Java written in Rust, designed to speed up development workflows involving AI coding agents like Claude Code. Traditional Gradle builds can take minutes, often delaying feedback for AI agents. Grounds operates by parsing code with tree-sitter to provide near-instant feedback on compile breaks, dependency issues, and dataflow errors without requiring a full JVM build. By caching stubs and analyzing only changed files, the tool typically provides results in under 0.3 seconds. While currently limited to Apple Silicon and lacking support for Kotlin, the project aims to bridge the gap between rapid iteration and code correctness. The tool is open-source and available on GitHub, offering a lightweight alternative for developers who need fast, reliable checks during the coding process.