Audit and verify AI-generated code in VS Code before cleanup, refactor, or launch.
This is the public profile for RomantiCode and our flagship product, LegacyDoc AI — a VS Code extension that turns real-world codebases (legacy projects, AI-generated MVPs, vibe-coded prototypes, inherited apps) into documented, audit-ready projects.
Not an open-source repository. This repository contains no product source code. Source code for the commercial VS Code extension is not published here. This is a public product profile and documentation entry point — it exists to give Google, users, and developer-tool directories a stable, citable URL for the product. The extension is distributed only via the VS Code Marketplace.
LegacyDoc AI generates, from your local codebase inside VS Code:
- Inline JSDoc, comments, and explanatory docstrings
- Standalone Markdown documentation files
- Mermaid architecture diagrams
- Module and folder summaries
- Areas to inspect and cleanup priorities
- AI code verification scope for architecture, risky flows, generated assumptions, and review evidence
- An AI-ready context pack for Claude Code, Cursor, or Codex
It is built for the workflow developers actually face in 2026:
The AI-generated app runs. Now what do you clean up, what's safe to refactor, and what context do you hand to the next developer or AI tool?
- Website: https://www.romanticode.com/
- Product page: https://www.romanticode.com/legacydoc-ai/
- Pricing: https://www.romanticode.com/pricing/
- Vibe coding cleanup services prep: https://www.romanticode.com/services/vibe-coding-cleanup-services/
- AI App Launch Audit: https://www.romanticode.com/services/ai-app-launch-audit/
- VS Code Marketplace: https://marketplace.visualstudio.com/items?itemName=ruilegendoc.legacy-doc-ai
- Audit and verify AI-generated code in VS Code: https://www.romanticode.com/tools/ai-code-audit-report/
- AI code audit cost calculator: https://www.romanticode.com/tools/ai-code-audit-cost-calculator/
- LLM cost regression checker: https://www.romanticode.com/tools/llm-cost-regression-checker/
- AI quota tracker: https://www.romanticode.com/tools/ai-quota-tracker/
- MCP security scanner: https://www.romanticode.com/tools/mcp-security-scanner/
- npm scope compromise checker: https://www.romanticode.com/tools/npm-scope-compromise-checker/
- AI code audit report template: https://www.romanticode.com/templates/ai-code-audit-report-template/
- CodeGraph token savings calculator: https://www.romanticode.com/tools/codegraph-token-savings-calculator/
- Vibe code cleanup cost calculator: https://www.romanticode.com/tools/vibe-code-cleanup-cost-calculator/
- AGENTS.md generator for AI coding agents: https://www.romanticode.com/tools/agents-md-generator/
- PROJECT.md template for AI coding agents: https://www.romanticode.com/templates/project-md-template-for-ai-coding-agents/
- AI codebase map example: https://www.romanticode.com/examples/ai-codebase-map-example/
- Codebase to LLM file packer guide: https://www.romanticode.com/blog/codebase-to-llm-file-packer/
- Bolt app code audit guide: https://www.romanticode.com/blog/how-to-audit-bolt-app-code/
- Lovable app code audit guide: https://www.romanticode.com/blog/how-to-audit-lovable-app-code/
- AI generated code architecture audit: https://www.romanticode.com/blog/ai-generated-code-architecture-audit/
- Claude Code context file checklist: https://www.romanticode.com/blog/claude-code-context-file-checklist/
- Cursor rules vs AGENTS.md: https://www.romanticode.com/blog/cursor-rules-vs-agents-md/
- VS Code AI extension permissions checklist: https://www.romanticode.com/blog/vs-code-ai-extension-permissions-checklist/
- AI app launch audit request template: https://www.romanticode.com/templates/ai-app-launch-audit-request-template/
- Example audit report: https://www.romanticode.com/examples/ai-code-audit-report/
- AI codebase understanding context map: https://www.romanticode.com/use-cases/ai-codebase-context/
- AI codebase map example: https://www.romanticode.com/examples/ai-codebase-map-example/
- Codebase to LLM file packer guide: https://www.romanticode.com/blog/codebase-to-llm-file-packer/
- Claude Code context file checklist: https://www.romanticode.com/blog/claude-code-context-file-checklist/
- Cursor rules vs AGENTS.md: https://www.romanticode.com/blog/cursor-rules-vs-agents-md/
- VS Code AI extension permissions checklist: https://www.romanticode.com/blog/vs-code-ai-extension-permissions-checklist/
- MCP security scanner: https://www.romanticode.com/tools/mcp-security-scanner/
- npm scope compromise checker: https://www.romanticode.com/tools/npm-scope-compromise-checker/
- Vibe coding cleanup services prep: https://www.romanticode.com/services/vibe-coding-cleanup-services/
- Vibe code cleanup prep: https://www.romanticode.com/use-cases/vibe-code-cleanup/
- Vibe code cleanup checklist: https://www.romanticode.com/blog/vibe-code-cleanup-checklist/
- Vibe code cleanup brief template: https://www.romanticode.com/templates/vibe-code-cleanup-brief-template/
- AI app launch audit service: https://www.romanticode.com/services/ai-app-launch-audit/
- AI app launch audit request template: https://www.romanticode.com/templates/ai-app-launch-audit-request-template/
- Bolt app code audit guide: https://www.romanticode.com/blog/how-to-audit-bolt-app-code/
- Lovable app code audit guide: https://www.romanticode.com/blog/how-to-audit-lovable-app-code/
- AI generated code architecture audit: https://www.romanticode.com/blog/ai-generated-code-architecture-audit/
- Legacy code documentation: https://www.romanticode.com/use-cases/document-legacy-code/
- AI coding model comparison workflow hub: https://www.romanticode.com/ai-models/
- Composer 2.5 context pack guide: https://www.romanticode.com/ai-models/composer-2-5/
- Gemini 3.5 Flash coding context guide: https://www.romanticode.com/ai-models/gemini-3-5-flash/
- Gemini Omni prompt and context guide: https://www.romanticode.com/ai-models/gemini-omni/
- Best way to get AI to completely refactor frontend code: https://www.romanticode.com/blog/best-way-to-refactor-frontend-code-with-ai/
- VS Code extension security checklist: https://www.romanticode.com/blog/vs-code-extension-security-checklist/
- AI code audit checklist: https://www.romanticode.com/blog/ai-code-audit-checklist/
- AI app production readiness checklist: https://www.romanticode.com/blog/ai-generated-code-production-ready-checklist/
- Vibe coding cleanup services — prepare an AI-generated app with an audit report, architecture map, cleanup priorities, and a quote-ready handoff before hiring cleanup help.
- Vibe code cleanup — prepare the architecture map, risk notes, cleanup priorities, and quote pack before refactoring AI-generated code.
- Vibe code cleanup checklist — run a cleanup checklist before refactor, AI coding agent work, or a specialist quote.
- Vibe code cleanup brief template — copy a quote-ready brief for product context, architecture notes, risk areas, cleanup priorities, and verification.
- Code graph for coding agents — estimate token, tool-call, and context prep savings before adopting graph-based agent workflows.
- LLM cost regression checker — estimate whether a pull request, model switch, reviewer agent, or tool-output change could raise recurring AI coding cost.
- AI quota tracker — plan Claude, Codex, Gemini, Copilot, and coding-agent quota usage before a team runs out mid-task.
- MCP security scanner — review MCP server tools, file access, secrets, network egress, prompt injection risk, and agent trust before connecting AI coding workflows.
- npm scope compromise checker — review lockfiles, lifecycle scripts, registry sources, and scoped-package changes before AI agents or CI trust a dependency update.
- Audit AI code — generate an audit-ready context pack before cleanup, refactor, review, or launch.
- AI code audit cost calculator — estimate discovery hours, report scope, and review budget before asking for an AI-generated code audit quote.
- AI code audit report template — copy the report structure for architecture context, inspection areas, cleanup priorities, and AI-ready handoff notes.
- AI codebase understanding — create a codebase map and PROJECT.md-style brief for Claude Code, Cursor, Codex, and other AI coding tools.
- AI codebase map example — study a sample codebase map with entry points, module ownership, data flow, risk areas, and review boundaries.
- Codebase to LLM file packer — choose files, map repo context, remove secrets, and build a clean LLM-ready bundle before pasting code into Claude, ChatGPT, Gemini, Cursor, or Codex.
- How to audit Bolt app code — review routes, Supabase policies, secrets, payments, failure paths, and cleanup handoff before launching a Bolt-built app.
- How to audit Lovable app code — review routes, Supabase policies, secrets, payments, failure paths, and cleanup handoff before launching a Lovable-built app.
- AI generated code architecture audit — map entry points, ownership, data flow, generated assumptions, and cleanup risk before AI agents or developers refactor an AI-built app.
- PROJECT.md template for AI coding agents — copy a project context template for Codex, Claude Code, Cursor, Copilot, and other AI coding agents.
- Claude Code context file checklist — prepare CLAUDE.md, PROJECT.md, AGENTS.md, repo maps, commands, boundaries, and verification notes before a Claude Code session.
- Cursor rules vs AGENTS.md — decide which instructions belong in Cursor rules, AGENTS.md, PROJECT.md, and generated audit reports.
- VS Code AI extension permissions checklist — review workspace access, provider routing, terminals, secrets, VS Code profiles, and approval evidence before installing AI coding extensions.
- Vibe coding cleanup specialist — understand cost, scope, and handoff prep before hiring a cleanup specialist.
- AI app launch audit — request launch-readiness review before real users, demos, cleanup services, or production handoff.
- AI app launch audit request template — copy a request brief before asking for launch-readiness review, cleanup quote, investor demo check, or developer handoff.
- Legacy code documentation — generate JSDoc, Markdown docs, architecture maps, module summaries, and AI-ready handoff context before cleanup.
- AI code audit pricing — compare free and Pro options for audit-ready docs, cleanup handoff, and VS Code context packs.
- VS Code extension privacy policy — review local-first code handling, BYOK provider keys, no RomantiCode code storage, and no extension telemetry.
- VS Code extension terms of service — review license scope, activation, acceptable use, support boundaries, and AI output responsibility.
- AI coding model comparison — compare new coding models with the same context pack, task boundary, and review gate.
- Composer 2.5 — prepare a Cursor Composer 2.5 context pack before long-running coding tasks.
- Gemini 3.5 Flash — plan Gemini 3.5 Flash coding-agent work with repo context and review evidence.
- Gemini Omni — turn multimodal product inputs into a reviewable implementation brief.
- Awesome AI Code Review: https://github.com/kodustech/awesome-ai-code-review
- Awesome AI Startups: https://github.com/nowork-studio/awesome-ai-startups
- Awesome Vibe Coding Resources: https://github.com/acvnace/awesome-vibe-coding-resources
- Install the extension from the VS Code Marketplace.
- Set your own AI provider API key (Anthropic, OpenAI, Google, xAI, or a custom OpenAI-compatible endpoint).
- Right-click a file or folder → generate docs, Markdown, or a full project analysis.
- Review the output in a diff preview before anything is written to your files.
- Share the generated context with Claude Code, Cursor, Codex, or with a human reviewer.
- Free — 5 generations per day. Inline JSDoc, Markdown, Mermaid diagrams.
- Pro ($29 one-time) — Unlimited generations, project-level analysis, custom prompt templates, up to 3 devices.
Compare plans: https://www.romanticode.com/pricing/
Get Pro directly: https://www.creem.io/payment/prod_6A1crhhY1jAOyEPrIkJPuY
Runs inside VS Code + BYOK. No code storage or proxying by RomantiCode.
Your code is sent directly from your machine to the AI provider you configure (Anthropic, OpenAI, Google, xAI, or a custom endpoint). RomantiCode does not see, store, or proxy your code.
LegacyDoc AI does not:
- Replace a professional security audit
- Replace SAST, CI, dependency scanning, or compliance review
- Automatically fix code
- Guarantee production readiness
- Modify your code without a diff preview
It helps you understand, document, prioritize, and prepare the codebase before cleanup, refactor, review, or handoff.
- Email: contact@romanticode.com
- VS Code Marketplace Q&A: https://marketplace.visualstudio.com/items?itemName=ruilegendoc.legacy-doc-ai
- Website: https://www.romanticode.com/
- docs/links.md — Canonical URL list for submissions and sharing
- docs/submission-copy.md — Ready-to-use copy for directory submissions, Marketplace descriptions, and short social posts
© RomantiCode. All rights reserved.