marvkr/better-design

🎨 Open-source design MCP server + shadcn/ui registry — AI design systems for Claude Code, Cursor, Codex, GitHub Copilot & any MCP client. 31 brand-grade themes (Linear, Stripe, Vercel, Notion, Apple, Supabase, Figma…) + design tokens, UI principles & WCAG rules. Install any component with one command.

What it solves

Better Design bridges the gap between generic AI-generated code and professional-grade UI. It provides AI coding agents with the specific design context—such as semantic CSS tokens, component overrides, and accessibility rules—needed to build interfaces that follow established design systems rather than producing "AI-looking" generic layouts.

How it works

It functions as an MCP (Model Context Protocol) server that integrates with AI coding tools like Claude Code, Cursor, and GitHub Copilot. It operates in two modes:

  1. Design Intelligence: It provides the agent with active design principles and review checklists (e.g., WCAG accessibility, visual hierarchy) to critique and refine its own output.
  2. Scaffold: It allows the agent to pull from a library of 31 curated design systems (e.g., Linear, Stripe, Apple, Notion). The agent uses these to select appropriate tokens, icon sets, and component code to ensure the generated UI matches a specific professional aesthetic.

Who it’s for

This is for developers and designers who use AI coding agents to build web interfaces and want to ensure their generated code is production-ready, accessible, and visually consistent with high-quality design standards.

Highlights

  • Provides an MCP-compatible design-system registry for AI agents.
  • Includes 31 pre-built design systems compatible with shadcn/ui.
  • Offers automated UI review tools to check for accessibility (WCAG) and visual design consistency.
  • Enables semantic matching of design systems based on natural language prompts.
  • Supports seamless integration with major AI coding environments like Cursor, Claude Desktop, and VS Code.

Related

  • Project
    MiniMax-AI/MiniMax-M2.1

    MiniMax-M2.1 is an open-source language model focused on agentic capabilities such as coding, tool use, and long-horizon planning. It provides state-of-the-art performance, especially in multilingual software engineering tasks, and can be deployed locally with frameworks like SGLang and vLLM.

  • Project
    ArtemXTech/personal-os-skills

    Personal OS Skills is a set of Claude Code plugins for Obsidian that automate tasks like syncing meeting notes, analysing voice dictations, managing tasks, importing Notebook LM content, recalling prior chat context, and exporting Claude sessions. Install via Claude Code’s marketplace, then run the individual skills to keep AI‑generated output inside your Obsidian vault. MIT‑licensed, with videos, blog posts, and a Discord community for support.

  • Project
    codexstar69/bug-hunter

    Bug Hunter is an adversarial AI code-audit tool that uses a multi-agent system of hunters, skeptics, and referees to filter out false positives and provide evidence-backed security and behavioral bug reports.

  • Project
    waybarrios/opencode-power-pack

    A collection of 55 portable workflows and security-focused sandbox tools that enable consistent, safe, and specialized task execution across multiple AI coding agents.