JetBrains Open-Sources KotlinLLM: an LLM That Writes Persistent Kotlin Source at the Call Site Instead of Being Called at Runtime

The JetBrains Blog (JetBrains Research)·medium signalXBlueskyLinkedInCopy link

JetBrains Research released KotlinLLM under Apache 2.0 — an IntelliJ plugin introducing 'Smart macros' whose bodies are generated Kotlin source code, regenerated only as the application hits runtime scenarios the existing code doesn't cover. Once generated, the code runs as ordinary Kotlin with no plugin dependency, no inference latency and no per-call token cost. Reported evaluations: an adapted Spring Petclinic completed 24/24 application scenarios after macro evolution with a 100% hot-reload rate and ~1% runtime overhead, and a GitHub Issue Radar hit ~0.89 recall against ground-truth beginner labels across 30,000+ issues. The repo ships committed generated sources so you can read what the model actually produced.