On
Building a Brand Context Library for Claude Code: A 30-Minute Setup

A brand context library gives Claude your tone, design language, and ideal customer profile—ensuring every output feels authentically yours instead of generically AI. Here's how to build one in about 30 minutes.

Why Claude's Output Feels Generic (And How Context Fixes It)

If you've used Claude for content creation and felt the results were technically correct but lacked your distinctive voice, you're onto something real. Claude operates on whatever context you provide. Without explicit brand guidelines, it fills gaps with defaults: neutral tone, templated structure, safe positioning.

The solution isn't writing better prompts—it's building better information architecture.

A brand context library gives Claude a persistent reference point for everything that makes your brand distinctive: your voice profile, design tokens, ideal customer profile (ICP), and market positioning. Once in place, Claude's output starts carrying your fingerprint instead of adopting someone else's style.

This guide walks you through building that system in roughly 30 minutes.

What Actually Is a Brand Context Library?

Think of it as an onboarding document you'd hand to a new freelance writer or designer on day one. Except instead of a PDF that gets skimmed once and forgotten, it lives in your project folder where Claude reads it whenever needed.

In Claude Code, you can place context files in your project root and reference them explicitly in prompts, or use a CLAUDE.md file that Claude automatically treats as persistent context. Both approaches work—what matters is that information stays structured, specific, and consistent.

A complete brand context library typically includes:

  • Voice and tone documentation—how you write, what to avoid, examples of good and bad
  • Design tokens documentation—colors, typography, spacing, and other visual variables in a format Claude can reference
  • Positioning documentation—your ICP, value proposition, competitive differentiation, and message hierarchy
  • Example documentation—real samples showing your brand voice in action

Each file is plaintext or Markdown. No special formatting required. The goal is making implicit brand knowledge explicit enough that an AI (or a new team member) can act on it without guessing.

Step 1: Build Your Voice and Tone Documentation

Voice is the hardest part to get right, so start here.

Identify your core voice attributes

Pick 3 to 5 adjectives describing how your brand actually sounds—not how you want it perceived, but how your best writing genuinely reads. Be honest. "Professional" is nearly useless. "Direct and slightly dry, like a senior engineer explaining something" is workable.

For each attribute, write two examples: one showing the attribute done well, and one showing it missing the mark.

Example format:

VOICE ATTRIBUTE: Direct

GOOD: "You need an API key before this feature works."
BAD: "To continue this process, you'll need to ensure you've completed the necessary steps to configure your API key settings."

VOICE ATTRIBUTE: Technical precision

GOOD: "Agent runs on a 15-minute cron schedule."
BAD: "Agent runs periodically in the background."

This before/after structure works better than descriptive paragraphs because Claude can pattern-match against concrete examples.

Document what you never say

Brand voice is as much about exclusion as inclusion. List words, phrases, and sentence structures your brand avoids.

Common categories worth documenting:

  • Banned words—jargon you hate, language misaligned with your brand, competitor terminology
  • Banned structures—never use passive voice, never start with a question, never use bullet points for emotional content
  • Banned tone—don't flatter, don't overuse exclamation marks, don't write like a press release

If you already maintain a banned words list for your team, paste it in. If not, spend 5 minutes listing things that make you cringe when you see them in writing.

Set reading level and formality

Be specific about word choice. "Friendly but professional" is too vague. Try this instead:

READING LEVEL: Target 8th-10th grade comprehension (Hemingway app score around 70)
SENTENCE LENGTH: Average 14–18 words. Mix short, punchy sentences with slightly longer ones.
PARAGRAPH LENGTH: 2–4 sentences in body copy. Single sentence is fine.
PRONOUNS: Use "you" and "we." Avoid "one" and third-person constructions.

Claude responds well to numerical constraints. "Short paragraphs" is unclear. "2-4 sentences" is actionable.

Add real examples

Pick your 5-10 best existing pieces of content—an email, a landing page section, a product description, a social post. Paste them verbatim into a section labeled CANONICAL EXAMPLES.

Step 2: Convert Your Visual Identity Into Design Tokens

This step is where most people drop the ball. They add voice guidelines but forget that Claude often needs to generate code, write design specs, or describe image treatment—and without visual identity documentation, it defaults to generic choices.

Design tokens are the fix. They're just named variables for your visual decisions.

Colors

Document your palette using hex codes, with labels reflecting usage rather than color names:

COLORS:
--color-primary: #1A1A2E
--color-secondary: #16213E
--color-accent: #0F3460
--color-highlight: #E94560
--color-background: #FAFAFA
--color-text-primary: #1A1A2E
--color-text-secondary: #6B7280
--color-success: #10B981
--color-warning: #F59E0B
--color-error: #EF4444

If you have a dark mode palette, document it separately using a --dark prefix convention.

Also note usage rules:

COLOR USAGE RULES:
- Accent color reserved for call-to-action buttons only. Never use decoratively.
- Never place text directly on highlight color.
- Backgrounds must always be light in onboarding flows.

Typography

Document your typeface system with enough detail that Claude can generate accurate CSS or design specs:

TYPOGRAPHY:
Font families:
  --font-heading: 'Inter', sans-serif
  --font-body: 'Inter', sans-serif
  --font-mono: 'JetBrains Mono', monospace

Scale (desktop):
  --text-xs: 12px / 1.5
  --text-sm: 14px / 1.5
  --text-base: 16px / 1.6
  --text-lg: 18px / 1.5
  --text-xl: 20px / 1.4
  --text-2xl: 24px / 1.3
  --text-3xl: 30px / 1.2
  --text-4xl: 36px / 1.1

Weight usage:
  Headings: 700 (bold)
  Subheadings: 600 (semibold)
  Body: 400 (regular)
  Labels/captions: 500 (medium)

Spacing, radius, and shadows

These matter more than most people realize. Without them, Claude will generate components that look visually inconsistent with your actual product.

SPACING (base unit: 4px):
  --space-1: 4px
  --space-2: 8px
  --space-3: 12px
  --space-4: 16px
  --space-6: 24px
  --space-8: 32px
  --space-12: 48px
  --space-16: 64px

BORDER RADIUS:
  --radius-sm: 4px
  --radius-md: 8px
  --radius-lg: 12px
  --radius-full: 9999px

SHADOWS:
  --shadow-sm: 0 1px 2px rgba(0,0,0,0.05)
  --shadow-md: 0 4px 6px rgba(0,0,0,0.07)
  --shadow-lg: 0 10px 15px rgba(0,0,0,0.10)

Logo and asset usage rules

Claude can't see your logo, but it can follow guidelines about how to reference or describe it in text and code:

LOGO USAGE:
- Primary logo: horizontal lockup (logo + brand name)
- Icon-only version: use when width < 120px
- Minimum clearspace: equal to icon height on all sides
- Never place on busy backgrounds
- Acceptable backgrounds: white, --color-primary, --color-background
- Never stretch, rotate, or recolor

Step 3: Document Your Positioning and ICP

Positioning is where Claude tends to get generic fast. Without clear guidance, it writes for an imaginary average customer—not your actual buyer.

Define your ideal customer

Be specific. Not "B2B SaaS companies," but:

IDEAL CUSTOMER PROFILE:

PRIMARY PERSONA: "Operations-minded founder"
- Company size: 5–50 employees
- Industries: Professional services, agencies, or SaaS
- Role: Founder or Head of Operations
- Core pain: Spending 15+ hours/week on tasks that should be automated
- Tech comfort: Uses Notion, Airtable, Zapier. Not a developer.
- Their fear: Loss of quality control when scaling without hiring more people
- Their desire: Stop being the bottleneck without expanding headcount
- How they talk: Results-focused. Impatient with jargon. Skeptical of hype.

If you have multiple customer personas, document each one using the same structure. Then add a note about which persona is primary for different content types (e.g., "Landing page copy targets Persona A. Email sequences target Persona B").

Write your positioning statement

Use a structured format Claude can reference directly:

POSITIONING:

For [operations-heavy founders at small service businesses]
Who are struggling with [spending too much time on repetitive work that slows their growth]
[MindStudio] is a [no-code AI agent builder]
That enables [non-technical people to create automated workflows in under an hour]
Unlike [traditional automation tools like Zapier or Make]
We [connect AI reasoning to your actual business tools, so agents can handle complex multi-step tasks, not just simple triggers]

This format forces clarity. If you can't fill it in coherently, your positioning statement isn't ready—useful to know before sending it to Claude.

Add message hierarchy

Not every message deserves equal weight. Tell Claude which claims should lead and which to downplay:

MESSAGE HIERARCHY:

Tier 1 (lead with these):
- Build speed: most workflows take 15–60 minutes
- No-code: built for non-technical users
- Real AI reasoning: not just trigger-action rules

Tier 2 (supporting claims):
- 200+ AI models available
- 1,000+ integrations
- Used by teams at TikTok, Microsoft, Adobe

Tier 3 (mention when relevant):
- Free to start
- Custom JavaScript/Python support for technical users
- Local model support

CLAIMS TO AVOID:
- Never say "best" or "only"
- Never compare directly by name unless asked
- Don't lead with pricing

Competitive context

Give Claude enough context to write differentiated content without being dismissive:

COMPETITIVE CONTEXT:
We're often compared to Zapier, Make, and n8n.
Key differentiators to mention if asked:
- Those tools are trigger-action based. We're built for multi-step AI reasoning.
- They require separate AI subscriptions and API setup. We have it built in.
- We're easier to build with for non-technical users.
Tone when mentioning competitors: respectful, objective, never dismissive.

Step 4: Assemble Your Directory Structure

Now piece everything together. Here's a clean directory layout:

/brand-context/
  CLAUDE.md ← main index file Claude reads first
  voice-and-tone.md ← voice attributes, rules, examples
  design-tokens.md ← colors, typography, spacing, shadows
  positioning.md ← ICP, positioning statement, message hierarchy
  copy-examples.md ← 5–10 canonical content samples
  competitive-context.md ← competitive landscape and differentiation

Write a master CLAUDE.md file

This is the file Claude automatically picks up in Claude Code. Use it as an index telling Claude what's in each file and when to reference them:

# Brand Context Index

This directory contains brand guidelines for [Company Name].
Always reference these files before creating any content, copy, or UI.

## Files in this directory:

- `voice-and-tone.md` — How we write. Read this before creating any content.
- `design-tokens.md` — Visual variables. Use these for all UI design work.
- `positioning.md` — ICP, positioning, and message hierarchy. Reference when writing marketing content.
- `copy-examples.md` — Real examples of our brand voice in practice.
- `competitive-context.md` — How to handle competitor comparisons.

## Quick reference:

Primary brand color: #1A1A2E
Primary typeface: Inter
Primary CTA copy: "Start building free"
Brand voice in three words: Direct. Specific. Trustworthy.

Step 5: Load It Into Claude Code

Once your directory exists, using it is straightforward.

Method 1: Reference the folder in your system prompt

If you're building a workflow or agent using Claude, add an instruction to read your brand context folder into the system prompt:

You are a content assistant for [Brand Name]. Before generating any content,
read and apply the documents in /brand-context/. Always follow the voice profile,
use terminology from the design tokens, and write for the ICP defined in positioning.md.

Method 2: Use CLAUDE.md in your Claude Code project

Claude Code supports a CLAUDE.md file at your project root that automatically loads as context for every session. Reference your brand context directory inside it:

## Brand Context

For all content generation, apply the brand principles in /brand-context/.
Key files:
- Voice and tone: /brand-context/voice-and-tone.md
- Positioning and ICP: /brand-context/positioning.md
- Design tokens: /brand-context/design-tokens.md

This means every Claude Code session in that project starts with your brand context already loaded—no manual prompting required.

Method 3: Reference specific files in individual sessions

For ad-hoc needs, you can call out files explicitly when starting a session:

Before we start, please read:
- /brand-context/voice-and-tone.md
- /brand-context/positioning.md

Now, help me draft a product launch email for [feature].

Step 6: Test, Refine, and Maintain

Building the directory is 80% of the work. The remaining 20% is iteration.

Run calibration prompts

Before using your directory in real work, run test cases. Ask Claude to generate something it's never seen before—a product description, a short email, a UI microcopy snippet—and compare it against your canonical examples.

Ask yourself:

  • Does the tone match?
  • Are visual decisions consistent with your design tokens?
  • Is the messaging at the right hierarchical level?
  • What feels off?

Then update the relevant files to address gaps.

Add edge case rules as you discover them

Every time Claude does something that feels wrong, document a rule that could have prevented it. Over time, your directory becomes a knowledge base of every brand decision that's ever come up.

This is especially useful for:

  • Content types you haven't covered yet (FAQs, error messages, notification copy)
  • Voice edge cases (how to write about pricing, how to handle bad news)
  • Technical context (variable naming conventions, code comment style)

Update regularly

Brand guidelines change. When you refresh messaging or evolve your design system, update your context directory at the same time. A stale brand context directory is worse than none at all—Claude will confidently generate outdated results.

Set a quarterly reminder to review each file. Takes 20 minutes if you've been maintaining it, and it's absolutely worth it.

Description: Learn how to create a brand context directory that ensures Claude generates content aligned with your voice, design system, and positioning.

Related Articles