https://github.com/1999azzar/article_generator
A web-based AI-powered article and novel outline generator built with TypeScript and Cloudflare Workers. Uses Google's Gemini AI to create professional content based on user specifications.
https://github.com/1999azzar/article_generator
Last synced: 2 months ago
JSON representation
A web-based AI-powered article and novel outline generator built with TypeScript and Cloudflare Workers. Uses Google's Gemini AI to create professional content based on user specifications.
- Host: GitHub
- URL: https://github.com/1999azzar/article_generator
- Owner: 1999AZZAR
- License: mit
- Created: 2025-11-30T07:46:46.000Z (7 months ago)
- Default Branch: main
- Last Pushed: 2026-03-08T08:37:06.000Z (4 months ago)
- Last Synced: 2026-03-08T11:59:02.953Z (4 months ago)
- Language: TypeScript
- Homepage: https://article-generator.azzar.workers.dev/
- Size: 411 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Quill™ - AI Writing Assistant
**Quill™** is an enterprise-grade AI writing assistant built on Cloudflare Workers and Google Gemini 3 models. It generates professional articles, short stories, and news briefs with precise author style mimicry.
**Live Demo:** [https://article-generator.azzar.workers.dev](https://article-generator.azzar.workers.dev)
---
## Features
### Content Generation
- **Articles**: Comprehensive, research-quality content (1800-2000 words)
- **Short Stories**: Narrative fiction with character arcs (2500-3000 words)
- **News Articles**: In-depth journalistic reporting (1200-1800 words)
- **Short News**: Concise news briefs in 5W1H format (400-600 words)
- **Novel Outlines**: Full chapter breakdown & synopsis
- **Chapter Continuity**: Intelligent story flow that maintains narrative consistency across chapters (using optimized summaries)
### Author Style System
Advanced style mimicry with curated prompts for 21+ specific authors:
- **Classic & Literary**: Hemingway, Austen, Orwell, Toni Morrison, etc.
- **Fantasy & Sci-Fi**: Tolkien, G.R.R. Martin, Neil Gaiman, J.K. Rowling
- **Contemporary**: Stephen King, Murakami, Margaret Atwood
- **Indonesian Authors**: Pramoedya Ananta Toer, Dee Lestari, Andrea Hirata
- **Non-Fiction**: Yuval Noah Harari
### Technical Highlights
- **Prompt Optimization**: Compressed prompts (60% token reduction) for faster generation
- **Smart Context**: Novel chapters use optimized summaries (last 3 chapters only) to prevent token bloat
- **Circuit Breaker**: Removed for stateless edge deployment stability
- **Gemini 3 Powered**: Uses latest `gemini-3-flash-preview` & `gemini-3-pro-preview` models
### UI/UX Design
- **Premium Aesthetics**: Dark mode, Glassmorphism, Inter font
- **Interactive UI**: Smooth button animations, categorized dropdowns
- **Mobile Responsive**: Optimized for all devices
---
## Installation
### Prerequisites
- Node.js 18+
- Cloudflare Account
- Google Gemini API Key
### Local Development
1. **Clone the repository:**
```bash
git clone https://github.com/1999AZZAR/article_generator.git
cd article_generator
```
2. **Install dependencies:**
```bash
npm install
```
3. **Start the development server:**
```bash
npm run dev
```
4. **Build the project:**
```bash
npm run build
```
---
## Deployment
We use **Wrangler** for direct deployment (GitHub Actions workflow has been removed).
1. **Login to Cloudflare:**
```bash
npx wrangler login
```
2. **Set API Key (Secret):**
```bash
npx wrangler secret put GEMINI_API_KEY
```
3. **Deploy:**
```bash
npx wrangler deploy
```
---
## Project Structure
```bash
article_generator/
├── src/
│ ├── index.ts # Worker Entry Point
│ ├── handler.ts # API Route Handlers
│ ├── gemini.ts # AI Logic & Prompt Engineering
│ └── ui.ts # Frontend Components (HTML/CSS)
├── public/ # Static Assets
├── PROMPT_OPTIMIZATION.md # Prompt Engineering Documentation
├── wrangler.toml # Worker Configuration (Ignored in Git)
└── package.json # Project Dependencies
```
---
## Documentation
- [Prompt Optimization Guide](PROMPT_OPTIMIZATION.md) - Detailed breakdown of how we reduced tokens by 60% while improving quality.
- [Model Verification](MODEL_VERIFICATION.md) - Details on Gemini 3 model usage and validation.
---
## License
MIT License. Created by [Azzar Budiyanto](https://github.com/1999AZZAR).