william.ai

AI ghostwriting platform for LinkedIn and Twitter that eliminates AI slop through structural constraints and multi-layer quality control.

Features

• Voice Profiles: Create unique voice profiles with 20+ rules, reference posts, and brand colors
• 5 Voice Modes: Thought Leader, Storyteller, Educator, Provocateur, Community Builder
• 6-Stage Generation Pipeline: Multi-stage refinement ensures high-quality output
• Quality Gates: 100+ forbidden phrases, specificity checks, hook strength validation
• Platform Conversion: Automatic LinkedIn → Twitter thread conversion
• Carousel Generation: Template-based carousel creation with brand colors

Tech Stack

• Frontend: Next.js 14 (App Router), TypeScript, Tailwind CSS
• Components: shadcn/ui, Framer Motion
• AI: Google Gemini 3 Flash API
• Storage: localStorage (MVP)
• Deployment: Vercel

Getting Started

Prerequisites

• Node.js 18+
• Google Gemini API Key (Get one here)

Installation

1. Clone the repository: ```bash git clone https://github.com/yourusername/william-ai.git cd william-ai ```

2. Install dependencies: ```bash npm install ```

3. Create a .env.local file: ``` GEMINI_API_KEY=your_gemini_api_key_here ```

4. Run the development server: ```bash npm run dev ```

5. Open http://localhost:3000 in your browser.

Usage

Creating a Voice Profile

1. Go to "Voice Profile" in the navigation
2. Enter a name and select target platforms
3. Add 20+ voice rules (sentence patterns, forbidden words, signature phrases)
4. Add 3 reference posts with engagement metrics
5. Select brand colors for carousels
6. Save your profile

Generating Posts

1. Go to "Create Post" and choose a flow:

   • Personal Experience: Share a story or lesson
   • Pattern Recognition: Share observations across situations

2. Select a voice mode that matches your content style

3. Answer the interview questions with specific details

4. Submit for generation (takes 30-60 seconds)

5. Review, edit, and export your post

Quality Requirements

• Posts must score 85+ to pass quality gates
• Zero forbidden phrases allowed
• Minimum 3 specific details (numbers, names, dates)
• Strong hook with number, contrarian claim, or personal stake

Project Structure

``` william.ai/ ├── app/ # Next.js App Router pages │ ├── api/ # API routes │ ├── create/ # Post creation flows │ ├── dashboard/ # Main dashboard │ └── voice-profile/ # Profile management ├── components/ # React components │ ├── ui/ # shadcn/ui components │ ├── interview/ # Interview flow components │ ├── generation/ # Generation UI components │ └── dashboard/ # Dashboard components ├── lib/ # Utilities and logic │ ├── context/ # React contexts │ ├── guardrails/ # Quality validation │ ├── pipeline/ # Generation pipeline │ ├── prompts/ # AI prompts │ └── utils/ # Helper functions ├── data/ # Static data │ ├── voice-modes.ts # Voice mode definitions │ └── carousel-templates.ts └── hooks/ # Custom React hooks ```

Deployment

Deploy to Vercel

1. Push your code to GitHub

2. Connect your repository to Vercel

3. Add environment variable:

   • GEMINI_API_KEY: Your Gemini API key

4. Deploy!

License

MIT

Credits

Built with Cursor and Claude Opus 4.5