Yes. Beyond classical modeling, text and wiki tools (see question F-1), a newer category of tools uses LLM-based agents to review existing documentation directly against the arc42 requirements.
One example is arc42agentic, an open-source agent system for IDEs/tools that support agentic coding assistants (e.g. VS Code with GitHub Copilot, or via APM):
- Review agents check a complete (or changed) arc42 documentation section by section against the arc42 requirements, and additionally run cross-section consistency checks (e.g. quality goals vs. solution strategy, context vs. building blocks, risks vs. quality requirements) to surface contradictions between sections.
- The agents automatically detect different documentation layouts (multiple files per section, flat files, or a single combined file), so they can be applied to documentation that doesn’t follow one specific tooling convention.
As with any AI-assisted tooling: treat the output as a structured review, not as a replacement for a human architect’s judgement: The usual rules for reviewing LLM output apply.