bmad-document-project

To generate documentation for an existing codebase, run an automated workflow on the project root to analyze code and output structured Markdown and JSON artifacts including architecture diagrams, data models, and API contracts.

AI-assisted brownfield documentation is the automated process of analyzing an existing codebase to detect project structure and tech stack, generating a structured PRD-ready guide with architecture diagrams, data models, and API contracts for AI context.

Yes, you can document monorepos and multi-part architectures automatically. The workflow provides per-part documentation for individual components and identifies integration points between them, alongside generating a master navigation index.

Yes, the documentation workflow requires configuration in the project directory before execution. It uses this configuration alongside pattern detection to classify the project type and generate appropriate structured Markdown and JSON outputs.

The generated codebase documentation artifacts are output in Markdown and JSON formats. These files include strict frontmatter and templating rules to ensure consistency across the master index, architecture diagrams, API contracts, and deployment notes.

If codebase documentation generation is interrupted, the workflow can resume partial progress. This ensures the analysis of the codebase and generation of structured Markdown and JSON outputs can continue without restarting from the beginning.