Brand Builder

Do NOT proceed to Phase 2 until all 9 answers are received.

PHASE 2 — Strategy Generation (auto, no approval needed)

Generate and save 01-strategy/brand-dna.md:

Brand DNA Template

# [Brand Name] — Brand DNA

## Positioning
[One sentence: We help [audience] [achieve X] unlike [competitor] who [does Y]]

## Archetype
[Choose one: Innocent / Explorer / Sage / Hero / Outlaw / Magician / Regular Guy / Lover / Jester / Caregiver / Creator / Ruler]
[One sentence explaining why this archetype fits]

## Core Values (4)
1. **[Value]** — [What this means in practice for this brand]
2. **[Value]** — [What this means in practice for this brand]
3. **[Value]** — [What this means in practice for this brand]
4. **[Value]** — [What this means in practice for this brand]

## Brand Personality
- [Adjective]: [What this means — e.g., "Bold: We say what others won't"]
- [Adjective]: [...]
- [Adjective]: [...]

## Customer Profile
**Who:** [Demographics + mindset]
**They want:** [Core desire]
**They fear:** [Core anxiety]
**Decision driver:** [What makes them choose]

PHASE 3 — Moodboard Generation

Step 1: Generate 4 moodboard images

Use Replicate (Flux 1.1 Pro) as default. Fallback: DALL-E 3.

Replicate CLI:

python -c "
import replicate, base64, os
prompts = [
  '$PROMPT_1',
  '$PROMPT_2',
  '$PROMPT_3',
  '$PROMPT_4',
]
os.makedirs('brand-kit/$BRAND/02-visuals/moodboard', exist_ok=True)
for i, prompt in enumerate(prompts):
    output = replicate.run('black-forest-labs/flux-1.1-pro', input={'prompt': prompt, 'aspect_ratio': '4:5'})
    with open(f'brand-kit/$BRAND/02-visuals/moodboard/mood-{i+1}.png', 'wb') as f:
        f.write(output.read())
    print(f'Saved mood-{i+1}.png')
"

How to write the 4 moodboard prompts — each image represents a different facet of the brand:

Prompt formula per image:

[Specific scene], [brand adjectives], [material/texture], [lighting cheat code], [color palette description], editorial photography, no text, no logos

Lighting cheat codes by tone:

• Underground/edgy → "harsh neon backlight, deep shadows, grain"
• Warm/human → "golden hour window light, soft diffusion, 35mm film"
• Clean/professional → "studio softbox, white seamless, product photography"
• Luxury → "Leica M6, available light, minimal grain, muted tones"
• Playful → "Polaroid wash, saturated pastels, flat lay"

Step 2: HUMAN APPROVAL GATE — Moodboard

Show all 4 images and ask:

Here are 4 moodboard images for [Brand Name].

[image 1] [image 2] [image 3] [image 4]

Direction: ✅ approved / ❌ redo [which ones + why] / 🔄 adjust [what to change]

Do NOT proceed to Phase 4 until approved.

PHASE 4 — Logo Concept Generation

Step 1: Generate 3 logo directions

Use Ideogram v2 for logo concepts (best for text rendering + design).

Ideogram via API or ideogram.ai browser:

Generate 3 distinct visual directions:

Logo prompt formula:

Professional logo design for "[BRAND NAME]", [STYLE], [PERSONALITY ADJECTIVES], [COLOR PALETTE], clean vector style, white background, no gradients, no shadows, centered composition

Example prompts:

# Concept A
Professional wordmark logo for "Bakchich", bold geometric sans-serif, underground and trustworthy, deep navy and warm gold, white background, centered, no icons, clean vector

# Concept B
Professional logo with icon for "Bakchich", abstract geometric B symbol paired with clean text, modern and bold, deep navy #1a2744 and gold #d4a843, white background, vector style

# Concept C
Minimal abstract logomark for "Bakchich", geometric shape representing exchange/connection, single color deep navy, ultra-minimal, white background, scalable icon

Step 2: Generate variations of each concept

For each concept: generate light version + dark version

# After approval of concept direction, generate:
# primary (on white), reversed (white on dark), icon-only, horizontal lockup

Step 3: HUMAN APPROVAL GATE — Logo Concepts

Here are 3 logo concept directions for [Brand Name].

[Concept A — Wordmark]
[Concept B — Icon + Text]  
[Concept C — Abstract Mark]

Which direction(s) to develop? A / B / C / A+B / other feedback

Do NOT proceed to Phase 5 until direction chosen.

PHASE 5 — Color Palette

Derive from approved moodboard. Extract dominant + accent colors.

Palette structure

Save as 02-visuals/colors/palette.json:

{
  "primary": "#HEX",
  "secondary": "#HEX",
  "accent": "#HEX",
  "neutral_dark": "#HEX",
  "neutral_light": "#HEX",
  "usage": {
    "primary": "headlines, key UI, logo",
    "secondary": "supporting elements, subheadings",
    "accent": "CTAs, highlights, links",
    "neutral_dark": "body text",
    "neutral_light": "backgrounds, containers"
  }
}

Generate a palette visualization image showing all 5 swatches with hex codes.

HUMAN APPROVAL GATE — Colors

Proposed color palette for [Brand Name]:

[palette visualization]

Primary: [name] [hex]
Secondary: [name] [hex]
Accent: [name] [hex]

Approve / adjust [which + new direction]?

PHASE 6 — Documentation Assembly

Once all visuals approved, generate the final 4 docs:

1. 01-strategy/brand-voice.md — AI-Ready Voice Document

# [Brand Name] — Brand Voice Document (AI-Ready)

## Role
You are a [ROLE] for [Brand Name], a [description] serving [audience].

## Voice Snapshot
- **[Adjective 1]:** [What this means — e.g., "Bold: Say it once, clearly. No hedging."]
- **[Adjective 2]:** [...]
- **[Adjective 3]:** [...]

## Writing Rules
- Tone: [formal/casual scale description]
- POV: Second person ("you")
- Sentence max: [X] words
- Contractions: yes/no
- Oxford comma: yes/no

## Vocabulary
**Always use:** [signature terms, brand phrases]
**Never use:** [banned words, AI clichés: "unlock", "leverage", "dive into", "skyrocket", "In a world where", "game-changing"]

## Situational Shifts
- Marketing/sales: [tone description]
- Customer support: [tone description]
- Crisis/complaints: [tone description]

## Gold Standard Examples (copy these patterns)
1. "[Best performing past copy example]"
2. "[Another example]"
3. "[Another example]"

2. 03-guidelines/visual-guidelines.md

Cover: logo usage rules, color usage, typography recommendations, spacing, do's/don'ts.

3. 04-ai-ready/master-copy-prompt.md

Pre-filled master prompt template using brand voice doc. Ready to paste into any AI tool.

4. 04-ai-ready/master-illustration-prompt.md

Brand illustration prompt skeleton with:

• Approved color hex codes filled in
• Moodboard reference URLs as --sref examples
• Style code / personalization hints
• Banned elements list

PHASE 7 — Final Kit Delivery

Brand kit for [Brand Name] is complete.

📁 brand-kit/[brand-name]/
  ✅ Brand DNA + Positioning
  ✅ AI-Ready Voice Document
  ✅ Moodboard (4 images)
  ✅ Logo Concepts ([A/B/C] direction)
  ✅ Color Palette (5 colors)
  ✅ Visual Guidelines
  ✅ Master Copy Prompt
  ✅ Master Illustration Prompt

Next steps:
- Refine logos in Figma/Illustrator using AI concepts as reference
- Run /writing-humanizer on all copy before publishing
- Use master-illustration-prompt.md as base for all future AI visuals

Tools & Fallbacks

Replicate setup:

pip install replicate
export REPLICATE_API_TOKEN=your_token

DALL-E 3 fallback:

from openai import OpenAI
client = OpenAI()
response = client.images.generate(model="dall-e-3", prompt=PROMPT, size="1024x1024", quality="standard")

Common Mistakes