What it does
LegacyDoc AI is a Visual Studio Code extension for documenting codebases that lack docs, such as inherited projects, legacy code and AI-generated apps. From the context menu on a file or folder you pick the output type, check the preview in the sidebar, and then apply it inline or save it as Markdown. Outputs include inline JSDoc and comments, standalone Markdown documentation, Mermaid diagrams of module dependencies and call flows, and a project context pack meant to be shared with tools such as Claude Code, Cursor or Codex. It also produces cleanup-readiness notes that point out areas to inspect and documentation that is missing. It is bring-your-own-key. You choose the provider yourself, from Claude, GPT, Gemini and Grok to any endpoint that follows the OpenAI API format, and the site says your code goes only to the provider you choose, not to RomantiCode servers. The site states that it documents rather than rewrites code, that it is not a formal security audit, and that generated documentation needs human review. JavaScript, TypeScript and Python are supported out of the box.
Capabilities
- Inserts JSDoc and explanatory comments while leaving code logic untouched
- Markdown documentation with function signatures, parameter descriptions and usage examples
- Mermaid diagrams of module dependencies and call flows
- AI-ready project context pack for coding tools
- Runs on your own API key, with Claude, GPT, Gemini, Grok or a compatible OpenAI-style endpoint
When to use it
- Documenting an inherited or undocumented codebase before changing it
- Preparing a project brief for an AI coding agent or a new developer
- Mapping architecture ahead of a refactor, cleanup or handoff
