Aditya Raj

DesignMD - Turn any website into an AI-ready design system

DesignMD analyzes live websites and extracts structured design intelligence including typography, spacing, colors, motion systems, breakpoints, and AI-ready prompts. Built for designers, frontend engineers, and AI builders who want to understand, document, and recreate production-grade interfaces faster. Paste any URL and generate actionable design system insights instantly.

Add a comment

Replies

Best
Aditya Raj
Hey everyone 👋 I built DesignMD after repeatedly struggling to analyze and recreate modern UI systems efficiently. Most inspiration tools focus only on screenshots, but I wanted something that could inspect deeper behavioral patterns from real websites. DesignMD extracts: typography systems spacing tokens color systems motion/animation behavior responsive breakpoints AI-ready prompts directly from live websites. The goal is to help designers, frontend engineers, and AI builders understand production-grade interfaces faster and turn them into structured workflows. Would genuinely love feedback, feature ideas, and edge cases to test 🙌
George Sostak

Good jo @adityarajdigital , that's exactly what I needed today!

George Sostak

@adityarajdigital What about the inside-app UI? I mean, usually, the design of the landing page is slightly different and lacks dynamic elements compared with the app's functional UI. All those popups, cards, lists, dashboards, banners, navigation, pills, you name it. Is it possible in your solution to supply multiple pages? Kind of walkthrough the app's multiple pages to catch the most of elements. I own an dev agency, and built 30+ apps, and the difference in public-facing pages with "logged-in kitchen" is always the case.

Aditya Raj

@gsostak That’s actually a very valid point, and honestly one of the biggest gaps I noticed myself while working on this.

Most existing solutions mainly focus on public-facing landing pages, but real products live inside the application layer — dashboards, flows, modals, states, navigation systems, empty states, onboarding, data-heavy views, etc. That’s where actual UX complexity exists.

My long-term vision is to gradually support deeper multi-page understanding and design extraction, not just static marketing pages. I’m still at the beginning and continuously improving things step by step from real user feedback like this.

Right now I’m trying to build it carefully instead of overpromising features early. But yes, internal app UI/UX systems, design consistency, and dynamic component understanding are definitely part of the direction I’m thinking about.

Really appreciate you bringing this up thoughtfully — feedback like this genuinely helps shape the product better.

George Sostak

@adityarajdigital thanks for considering this improvement, as at this exact moment, I'm in process of creation design.md file for the new app and 'internal' UI is 80% of the design.md

Aditya Raj

@gsostak Yeah, really appreciate you bringing this up honestly 🤝

I’m still building/improving things step by step, and as a solo builder it takes time organizing and exploring all these directions properly 😅

But internal product UI understanding is definitely something I want to go deeper into.

Rivra

Retrofitting existing sites for AI readiness is a huge pain point. Does DesignMD automatically map existing CSS variables to modern design system tokens?

Aditya Raj

@rivra_dev Yes — that’s one of the core goals behind DesignMD.

The system analyzes live DOM/CSSOM structure and attempts to extract existing frontend primitives including CSS variables, spacing scales, typography systems, color tokens, breakpoints, and interaction patterns directly from production implementations.

Instead of treating the UI as static screenshots, DesignMD tries to map real implementation details into a more structured, reusable design-system-oriented format that can be used for documentation, recreation, or AI-assisted workflows.

There’s still a lot we’re improving around token normalization and semantic mapping, but production CSS variable extraction and token grouping are already a major part of the pipeline.