Transform meeting transcripts into professional, bespoke brand identities. Building general design intelligence.
Marque is an AI-powered logo generator that analyzes meeting transcripts and creates professional logo concepts. The output is designed to look like thoughtful, human-crafted design work—not generic AI clip-art. Each logo feels as if a human designer worked directly with the client, translating their vision into visual form.
- Transcript Analysis: Deep extraction of brand values, industry, audience, and tonal "vibe"
- Intelligent Design Decisions: AI determines optimal logo type (wordmark, pictorial, or abstract)
- Ranked Concepts: Three strategically different approaches, ordered from best (1) to least optimal (3)
- Design Rationale: Every logo includes objective, third-person documentation explaining geometric and conceptual choices
- True Vector SVG Export: Download logos as production-ready SVG with smooth bezier curves via ImageTracer vectorization
- Professional UI: Lightfield-inspired utilitarian interface with "high-end industrial laboratory" aesthetic
- Edge Case Handling: Graceful fallbacks when company names aren't explicitly mentioned
| Type | Description | Best For |
|---|---|---|
| Wordmark | Text-based logo using company name | Brands with distinctive names |
| Pictorial | Recognizable real-world object (like Apple's apple) | Clarity and name reinforcement |
| Abstract | Geometric shapes conveying symbolic meaning (like Nike's swoosh) | Flexibility and trademark uniqueness |
- Framework: Next.js 15 (App Router)
- Styling: Tailwind CSS v4
- AI: Gemini 2.0 Flash via Google AI API
- Database: Supabase (optional, for persistence)
- Deployment: Vercel
- Node.js 18+
- npm or yarn
- Gemini API key
# Clone the repository
git clone https://github.com/yourusername/marque-logo-generator.git
cd marque-logo-generator
# Install dependencies
npm install
# Set up environment variables
cp .env.example .env.local
# Edit .env.local and add your API keysCreate a .env.local file with:
# Required
GEMINI_API_KEY=your_gemini_api_key_here
# Optional (for persistence)
NEXT_PUBLIC_SUPABASE_URL=https://xxx.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=eyJ...npm run devOpen http://localhost:3000 to see the app.
If you want to persist generations, create this table in Supabase:
CREATE TABLE generations (
id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
transcript TEXT NOT NULL,
logo_data JSONB NOT NULL,
created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
-- Allow public access (no auth required)
ALTER TABLE generations ENABLE ROW LEVEL SECURITY;
CREATE POLICY "Allow public access" ON generations
FOR ALL USING (true) WITH CHECK (true);- Push your code to GitHub
- Import the project to Vercel
- Add environment variables in Vercel's project settings
- Deploy!
The logo generator follows these principles of good logo design:
- Simple – Easy to recognize and remember at a glance
- Distinctive – Clearly different from competitors; avoids clichés
- Relevant – Fits the brand's purpose, audience, and tone
- Scalable – Legible at any size, from favicon to billboard
- Versatile – Functions in black & white and across mediums
- Timeless – Avoids trends that date quickly
- Clear – Communicates without explanation or excess detail
├── app/
│ ├── api/
│ │ └── generate/
│ │ └── route.ts # Logo generation endpoint
│ ├── layout.tsx # Root layout with fonts
│ ├── page.tsx # Main dashboard
│ └── globals.css # Global styles
├── components/
│ ├── LogoCard.tsx # Individual logo display with export
│ ├── LogoRenderer.tsx # SVG string → viewable/scalable image
│ ├── ResultGallery.tsx # Ranked logo results grid
│ └── TranscriptInput.tsx # Transcript input form
├── lib/
│ ├── prompts.ts # AI prompt engineering (design principles, constraints)
│ ├── supabase.ts # Database client
│ └── transcriptParser.ts # Transcript analysis utilities
└── .env.local # Environment variables
All generated logos adhere to these constraints for professional quality:
- Monochrome — Black on white only, enforcing visual clarity
- Geometric simplicity — Limited to
<path>,<circle>,<rect>,<polygon> - Stroke consistency — Uniform weight across all elements
- Self-critique — AI evaluates output against design principles before finalizing
Logos are generated as raster images by Gemini's image generation model, then converted to true vector SVG using ImageTracer bitmap-to-vector tracing. This produces clean SVG paths with smooth bezier curves—identical to professionally designed logos—rather than pixelated rectangle approximations. The exported SVG scales infinitely without quality loss.
See PROCESS.md for detailed development process documentation, including prompt engineering iterations and design decisions.