Generate type-safe APIs, validate architecture rules, and make your BPMN models AI-ready — straight from your build.
Your Zeebe and Camunda code is full of references to BPMN elements — process IDs, message names, task types. Rename one in the modeler and nothing warns you until it fails in production. bpmn-to-code generates a type-safe Process API from your .bpmn files via Gradle, Maven, or the in-browser tool.
Missing implementations, undefined timers, inconsistent naming — without architectural tests, BPMN quality drifts silently. The testing module ships as a JUnit-friendly rule set you drop into any project. Catch violations in your test suite, not in staging.
See the validation rules →Raw BPMN is 80% rendering metadata — shapes, bounds, waypoints. Agents waste context on pixels instead of process logic. bpmn-to-code emits a semantic JSON model built for AI context, not for diagrams, and exposes it over an MCP server your agent can query directly.
Read the JSON surface spec →Drop-in agent skills, built on everything bpmn-to-code offers. Integrate the plugin into your project in one prompt, scaffold process services from a diagram, and get architectural tests written for you — while you focus on modeling the business logic. Works with Claude Code and any other agent supporting the agentskills.io standard.
See all skills →Create a complete Spring Boot process service from a BPMN file — hexagonal layers, workers, Docker Compose, all wired to the generated Process API.
Build a project-specific BPMN style guide with an interactive wizard, then let the agent validate every new diagram against it.