Get started / Introduction

Introduction

Brandbook keeps one brand system per brand and makes it something your tools can read, follow and check against. These docs are for connecting those tools.

The brand system

Every brand starts from a scan of its website and is edited by people in the app. However you reach it, you get the same document, in five parts:

LogoVersions, minimum size, clear space and don’ts.

ColourHex values with names, roles and contrast.

TypeHeadline and body faces.

VoiceTraits, words to use and avoid, rules and example headlines.

ImageryPhotos, a description and rules.

Every change bumps the brand’s version and is logged with where it came from, including changes your tools make through the API.

Three ways in

MCP serverFor agents. Claude Code, Claude Desktop, Cursor and VS Code read your brand and check their own work.REST APIFor your code. Fetch the brand, check drafts, add rules.Plain filesFor everything else. brand.md for prompts and projects, tokens.json for Figma and code.

The MCP server and the REST API use the same key. A key belongs to a workspace, so one key covers every brand in it.

Checks

A check takes copy, a design or a list of colours and returns a score out of 100, a verdict and findings, most with a fix. Every check is logged on the brand’s Check page, so you can see what your tools have been sending.

With or without AI

Checks run on a rule engine first. With AI turned on, a model review adds findings and reads images; it never overrules the rules. Without AI, every endpoint still answers from the rules.

Quickstart →