shadcn-ui-mcp-server and heroui-mcp
These are competitors: both provide MCP servers that give LLMs context about React component libraries and their usage, allowing developers to choose between shadcn/ui's headless, Tailwind-based components or HeroUI's pre-styled component system.
About shadcn-ui-mcp-server
Jpisnice/shadcn-ui-mcp-server
A mcp server to allow LLMS gain context about shadcn ui component structure,usage and installation,compaitable with react,svelte 5,vue & React Native
Implements the Model Context Protocol (MCP) with direct GitHub API access to shadcn/ui repositories, providing AI assistants with real-time component source code, demos, and block implementations. Supports dual transport modes—stdio for CLI and SSE for HTTP-based multi-client deployments—with intelligent caching and configurable GitHub token authentication for rate limit optimization. Enables framework-agnostic component integration with switchable primitive libraries (Radix UI or Base UI) and one-click installation via `.mcpb` bundles for Claude Desktop.
About heroui-mcp
heroui-inc/heroui-mcp
⚙️ MCP servers for HeroUI React v3 and HeroUI Native.
Implements dual MCP servers exposing HeroUI component documentation, theme tokens, and source code as context resources for AI assistants like Claude and Cursor. Each server (React and React Native variants) exposes searchable component discovery, props/types, usage examples, and design tokens via standardized MCP protocol endpoints. Built as a monorepo with Cloudflare Workers backends for automatic data extraction from upstream HeroUI packages, keeping AI context synchronized with latest component APIs.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work