Labsco
MCP SERVER

Jakarta Migration MCP

by adrianmikula

See what a javax-to-jakarta upgrade will break before you touch the code, then let OpenRewrite do the rename.

Code Generation, Scaffolding & Migration
Summary
Migration work you can read before you agree to it.

Most javax-to-jakarta upgrades go wrong in the order of operations: the refactor runs first and the blocker turns up later, in a dependency nobody can rebuild. This puts the measurement first — readiness, blockers, versions, a plan — and only then applies the automated rewrite, with the assistant chaining those steps out of a conversation instead of you working through a scan report tab by tab.

What it is

An IntelliJ IDEA plugin that hands its Java EE to Jakarta EE migration analysis to the JetBrains AI Assistant as MCP tools. The plugin does the static analysis over source, dependencies and configuration; the assistant can then ask it for a readiness score, a blocker list filtered by severity, compatible version recommendations or a phased plan, apply the OpenRewrite recipes that rewrite javax packages to their jakarta equivalents, and re-run the checks to confirm the result compiles and passes its tests. The tools are served inside the IDE, so this is the JetBrains assistant's surface rather than a server you register with a desktop client. Risk analysis, dependency scans, version recommendations and strategy insights are in the free tier; one-click and sandboxed refactoring, platform detection, advanced scans, PDF reports and the AI-assisted migration tools are on a paid subscription.

What you get
  • A readiness score for a project, with the dependencies it is blocked on and what to do about each
  • An impact report covering affected dependencies, breaking changes, estimated effort and a risk assessment
  • The blockers on their own, filtered by severity, when you only want the things that stop a build
  • Jakarta-compatible versions recommended for the dependencies you already have, with alternatives and upgrade paths
  • OpenRewrite refactoring applied across file patterns you choose, with a dry run before you commit to it
  • A compiled JAR scanned for incompatible classes, for dependencies you cannot rebuild from source
  • One dependency moved to a recommended version, rather than the whole tree at once
  • A phased migration plan with tasks, milestones and a rollback path
  • A validation pass that compiles the project and runs its tests, and reports whether the migration actually landed