Pattern MCP helps coding agents make better UI decisions. It checks real components from shadcn/ui and 21st.dev against what the product needs. If nothing fits, Pattern finds a real reference from Mob
Pattern is an MCP server that helps coding agents make better UI component decisions.
See Quick Start below to add your Anthropic API key and connect Pattern to your MCP client.
Instead of returning a list of search results, Pattern looks at what you need, checks real components against that need, and tells the agent whether to:
- Use an existing component from shadcn/ui, 21st.dev, or ReUI - Build a custom component, using a real product reference from Mobbin and/or Figma Community
Pattern is designed for agents to use while they are building.
- recommendcomponent — evaluates a UI component need and returns a structured recommendation. - extractrequirements — runs just the requirement-extraction step on its own, so you can inspect or hand-edit the checklist before recommendcomponent spends its search+score budget on it. - recordcomponentdecision — records what the agent actually did so future recommendations in the same project can take that decision into account.
For each recommendcomponent call, Pattern:
1. Checks whether the need is a simple primitive that doesn't require a search. 2. Turns the request into a set of specific requirements, unless a checklist was already supplied (see checklist). 3. Searches for matching shadcn/ui, 21st.dev, and ReUI components. 4. Checks each candidate against the requirements using evidence from the actual component. 5. Calculates how much of the requirement is covered. 6. Decides whether to use an existing component or build a custom one. 7. If a custom build is needed, searches Mobbin and Figma Community for real product example
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/donaldrichard19-LVD--pattern-mcp.html)
A Model Context Protocol server for persistent project-specific AI agent context.
mkpoli/gmail-mcpMulti-account Gmail MCP server on Cloudflare Workers — one deployment, one Google sign-in per connection, any MCP client
uidxkit/uidxA Visual Design that ships with your code. Agents and humans work together on the same canvas, where every change is a diffable JSX stored in a .uidx file inside your repo. Built on OpenPencil and Ski
metahub-ai/assayAn open, reproducible trust layer for AI artifacts — skills, MCP servers, agents, and plugins. Assay inspects code, runs sandboxed behavioral audits, and generates verifiable reports. Help us build th
motif-Labs/motifCoding agents are single-player: each session starts cold, in its own tool, blind to what the rest of the team already worked out. Motif makes your team's multiplayer — every Claude Code, Codex and Cu
syoooo/figma-relaiYour AI, on the canvas — connect Claude Code, Cursor, or any MCP client to Figma: read, edit, audit, build. The file itself carries your rules and rulings. Works on every Figma plan.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.