Brandrail

au.com.brandrailPublished by brandrail.com.auv0.10.1Updated Nov 7, 2026

Design system guardrails for AI coding tools: tokens, components and a brand checker. Runs locally.

VerifiedSTDIODesktop onlyDeveloper ToolsMedia & Design

Overview

AI-generated overview

Local MCP server that gives AI coding tools your design-system rules, tokens and components, and checks generated UI code against them.

What you can do
  • Read the project's design-system rules and the tokens (colour, spacing, radius, type) with their CSS variables
  • Look up components by need or by name, with imports, props, rules and examples for the project's framework
  • Fetch whole page templates (list, detail, form, settings, dashboard, sign-in) built from your components
  • Get a migration map showing what replaces Material, MUI, Bootstrap, Ant, Chakra or PrimeNG pieces
  • Validate generated snippets and get a brand score with per-issue fixes
When to use it
Useful when an assistant writes front-end UI in a repo that follows a Brandrail Brand Kit or a defined design system, and you want generated markup to use approved components and tokens instead of ad-hoc styles. Also handy when migrating away from a third-party UI library.
Requirements
Runs locally over stdio, typically started with npx brandrail mcp; Node.js and npm are needed. Works without an account using starter tokens; an optional BRANDRAIL_KEY environment variable (or a saved sign-in from the connect command) loads your organisation's Brand Kit. Desktop only.
Before you install
The optional BRANDRAIL_KEY is a secret; the connect flow stores an API key in a local credentials file and only its hash is sent to Brandrail. Code is checked locally and not uploaded, but the report option sends scores, and the upload flags on the CLI send results to the dashboard. The scan command drives a real browser and can use a stored auth state file for pages behind a login.

Installation

In SourceWeft

  1. Open Brandrail in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

brandrail

Keep AI-generated UI on-brand. brandrail checks code and live pages against your design system (Brand Kit), and writes your rules into the files AI coding tools read.

bash
npx brandrail init            # brandrail.config.json, src/brandrail.tokens.css + AGENTS.md, CLAUDE.md, Cursor/Copilot/Windsurf/Gemini rulesnpx brandrail check           # violations + brand score for this reponpx brandrail scan <url>      # check a running page in a real browser

With a Brandrail account

bash
npx brandrail connect                     # sign in through the browser; sets up the Brand Kit, AI rule files,                                          # the MCP server for Claude Code / Cursor / Copilot, and the UI kitnpx brandrail check --changed --upload    # only changed lines, results in the dashboard

connect makes an API key on your machine and saves it in ~/.config/brandrail/credentials.json (mode 600); Brandrail only ever receives its hash, and the files it writes into the repo hold no key. AI tools start npx brandrail mcp, which checks code on your machine: code is never uploaded. In CI, set BRANDRAIL_KEY instead.

MCP server

brandrail mcp is an MCP server for Claude Code, Cursor, Copilot, Codex, Windsurf and any other MCP client. It runs on your machine over stdio; code is checked locally and never uploaded. It works without an account, using starter tokens; connect an account (or set BRANDRAIL_KEY) to use your own Brand Kit.

json
{  "mcpServers": {    "brandrail": { "command": "npx", "args": ["-y", "brandrail", "mcp"] }  }}

npx brandrail connect writes this config for you, plus the rule files each AI tool reads.

ToolWhat it does
get_guidelinesThe project's design system rules: components to use in its framework, allowed tokens, what the checker enforces. Called once before writing UI.
get_design_tokensEvery colour, spacing, radius and type token with the CSS variable to use.
find_componentFinds the component for a need ("date picker", "primary action") with its import, rules and an example.
get_component_docsFull docs for one component in the project's framework: import or tag, props, rules, example.
get_templateA whole page (list, detail, form, settings, dashboard, sign-in) built from your components, with the files to create and the page's rules.
get_migration_mapWhat replaces each piece of Material, MUI, Bootstrap, Ant, Chakra or PrimeNG.
validate_snippetRuns the checker on generated code and returns a brand score and each fix.

All tools are read-only. Docs: https://www.brandrail.com.au/docs/connect

Commands

Command
check [paths]--format pretty|json|markdown|github, --fail-below 80, --changed [base], --upload
scan <url>--pages 3, --storage-state auth.json (pages behind a login), --wait 2000, --upload. Needs npm i -D playwright && npx playwright install chromium
init--from-api, --mcp, --tokens-css <file>, --no-rules, --force. Writes the Brand Kit as CSS variables; import that file once at your app entry
import <tokens.json>DTCG, Tokens Studio or Figma variables JSON into the config. --replace, --primitives
rulesList the rules
connect--no-browser, --no-install, --remote (hosted MCP server), --framework, --force (sign in again)
mcpThe MCP server over stdio, for AI tools. --framework, --report (send scores, never code)
migrate [paths]What's left on the way to the design system: other UI libraries (Material, MUI, Bootstrap, Ant Design, Chakra, PrimeNG) per library and per file. --json, --top 30, --upload
reportA shareable HTML handover report (paths and counts, never code). --out file.html
whoami, logoutShow or remove the saved sign-in

Nothing is uploaded unless you pass --upload, and code snippets are never uploaded.

Licence

MIT. The Brandrail web app and dashboard are a separate, paid service: https://www.brandrail.com.au/pricing

Source: packages/cli/README.md at commit e498958

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v0.10.1LatestNov 7, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts