1AI-polish - AI Academic Writing Polishing System

1 minute read

Published:

An AI-powered academic writing polishing system using DeepSeek-R1 API, featuring multi-style text polishing, AI detection, and hybrid deployment architecture.

Key Features

  • Multi-Style Text Polishing: Academic, formal, casual, and creative writing styles
  • AI Detection: Analyze text for AI-generated content probability
  • Thinking Process Display: Show DeepSeek-R1’s reasoning process
  • Hybrid Deployment: GitHub Pages frontend + local FastAPI backend
  • Complete API: RESTful API with Swagger documentation

Tech Stack

  • Backend: FastAPI, Python 3.9+
  • AI Model: DeepSeek-R1 via Volcano Engine API
  • Frontend: HTML/CSS/JavaScript
  • Deployment: GitHub Pages + Local Server

Quick Start

Backend Setup

# Clone repository
git clone https://github.com/ktwu01/1AI-polish.git
cd 1AI-polish

# Create virtual environment
python -m venv fastapi_env
source fastapi_env/bin/activate

# Install dependencies
pip install -r requirements.txt

# Configure API key
echo "ARK_API_KEY=your-api-key-here" > .env

# Start server
uvicorn app.main_production:app --reload --host 0.0.0.0 --port 8000

Access Frontend

  • Online: https://ktwu01.github.io/1AI-polish/
  • Local: Open docs/index.html in browser

API Endpoints

Text Polishing

POST /api/v1/process
{
  "content": "Your text here",
  "style": "academic"
}

AI Detection

POST /api/v1/detect
{
  "content": "Text to analyze"
}

Supported Styles

StyleDescriptionUse Case
academicProfessional and rigorousResearch papers, academic writing
formalSolemn and decentBusiness documents, formal occasions
casualConcise and clearPopular science, general reading
creativeNovel and interestingCreative writing, marketing

Architecture

Hybrid Deployment:

  • Frontend hosted on GitHub Pages (free, fast)
  • Backend runs locally (API key security)
  • Frontend connects to local API automatically

Project Structure

1AI-polish/
├── app/                    # Backend application
│   ├── api/v1/            # API routes
│   ├── services/          # Business logic
│   └── models/            # Data models
├── docs/                  # Frontend files
├── tests/                 # Test files
└── requirements.txt       # Dependencies

Performance

  • Processing time: 10-30 seconds (DeepSeek-R1)
  • Supports batch processing
  • Real-time AI probability detection
  • Live Demo: https://ktwu01.github.io/1AI-polish/
  • GitHub: https://github.com/ktwu01/1AI-polish
  • API Docs: http://localhost:8000/docs (when running)

Made with 💖 for academic writers