Skip to main content
MCP server CX Codex

Brandrail — Claude Code MCP server: UI checks

The Brandrail MCP server for Claude Code gives an agent a project’s design rules, tokens, and component details. It can validate generated UI code against the brand system.

MIT ★ 3 updated 2026-11-13

Install into Claude Code · Platform CLI All methods and platforms ↓
claude mcp add brandrail -e BRANDRAIL_KEY=<BRANDRAIL_KEY> -- npx -y brandrail
Open on GitHub

What Brandrail does

Brandrail exposes a project’s design rules, tokens, and component documentation to an AI agent. The agent can look up a component for a task, retrieve a page template, or check a generated snippet and its required fixes. MCP tools are read-only.

The server runs locally with npx brandrail mcp; code snippets stay on the machine. Starter tokens work without an account. An account or BRANDRAIL_KEY connects the project’s own Brand Kit. npx brandrail connect configures MCP and writes rule files for AI tools.

Checking a live page with scan also requires playwright and Chromium. Results are sent only with --upload, and code snippets are never uploaded.

What it helps with

  • check generated UI against a project’s design rules
  • find a project component for a specific interface task
  • retrieve the CSS variable for a design token
  • map Material or MUI elements to project components

Example prompts for the agent

  • Check this UI snippet against the project’s design rules and list the fixes.
  • Find the component for a date picker and show how to use it.
  • Get a settings page template built from the project’s components.

How to install Brandrail

Platform CLI

claude mcp add brandrail -e BRANDRAIL_KEY=<BRANDRAIL_KEY> -- npx -y brandrail

Replace <…> with your own values

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about Brandrail

How do I install Brandrail in Claude Code?

Run claude mcp add brandrail -e BRANDRAIL_KEY=<BRANDRAIL_KEY> -- npx -y brandrail. Set BRANDRAIL_KEY to connect your own Brand Kit.

What does Brandrail do?

It provides an AI agent with design rules, tokens, and component documentation, then checks code against the project’s brand system. Code is checked locally.

What does Brandrail need to run?

MCP runs through npx. Starter tokens work without an account; an account or BRANDRAIL_KEY connects a custom Brand Kit. The Brandrail web app and dashboard are paid services.

Similar MCP servers

All MCP servers for Design →

Project sponsors

Companies that keep this analytics open