Tamil AI Proofreading Platforms

A web-based platform that enables Tamil writers to upload text for AI-assisted proofreading. The system uses a tiered model workflow: a lightweight model handles up to a word limits, and escalates to a deeper model for longer or complex texts. Users access services via secure login and pay per use or via .

Features

Core Feature

• User Authentication: Secure login via email/password, OTP, and social login (Google, Facebook)
• Text Submission: Upload or paste Tamil text for proofreading
• Tiered Model Workflow:
  • Model A: Handles up to 500 words (fast, basic grammar/spelling)
  • Model B: Handles 500+ words (deep syntax, semantic checks)
• Payment Integration: Stripe and Razorpay support for global and Indian users
• Pay-per-use: ₹10 per 500 words (Model A), ₹20 per 500 words (Model B)
• Subscription Plans: Monthly/Yearly tiers with usage limits
• Dashboard: Submission history, word usage tracker, payment status, model usage breakdown
• Admin Panel: User management, model performance logs, payment reconciliation, analytics

Tech Stack

Backend

• Go 1.21+ with Gin framework
• PostgreSQL with GORM
• JWT for authentication
• Stripe & Razorpay for payments
• OpenAI API for LLM integration

Frontend

• Express (Node.js) with EJS templates — deployed to Vercel
• Tailwind CSS for styling
• Axios for API calls; workspace and home editor in express-frontend/public/js/

Project Structure

tamil-proofreading-platform/
├── backend/                 # Go API server
│   ├── cmd/server/         # Application entry point
│   ├── internal/
│   │   ├── config/         # Configuration
│   │   ├── handlers/       # HTTP handlers
│   │   ├── middleware/     # Auth, CORS, rate limiting, sanitization
│   │   ├── models/         # Database models
│   │   └── services/       # Business logic
│   └── migrations/         # Database migrations
├── express-frontend/        # Express app (Vercel + Docker)
│   ├── views/              # EJS pages (home, workspace, drafts, tools)
│   ├── routes/             # API proxy, auth, index
│   ├── public/js/          # workspace.js, home-editor.js, auth-utils.js
│   └── server.js           # Entry point
└── docs/                   # Documentation

Local IME (Aksharamukha) Development

1. Start the Python runner

cd services/aksharamukha-runner
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn app:app --host 0.0.0.0 --port 8088

2. Start the Go backend with IME enabled

cd backend
IME_ENABLED=true IME_CACHE_ENABLED=true AKSHARA_URL=http://localhost:8088 go run cmd/server/main.go

3. Test runner directly

curl -X POST http://localhost:8088/transliterate \
  -H "Content-Type: application/json" \
  -d '{"text":"enathu","mode":"spoken"}'

4. Test IME suggest through backend

curl "http://localhost:8080/api/v1/ime/suggest?q=enathu&mode=spoken&limit=8"

Cloud Run (optional):

gcloud run deploy aksharamukha-runner \
  --source services/aksharamukha-runner \
  --region asia-south1 \
  --allow-unauthenticated

Getting Started

Prerequisites

• Go 1.21 or higher
• Node.js 18 or higher
• PostgreSQL 12 or higher
• OpenAI API key
• Stripe account (optional)
• Razorpay account (optional)

Backend Setup

1. Navigate to the backend directory:

cd backend

2. Install dependencies:

go mod download

3. Set up environment variables (copy .env.example to .env):

cp .env.example .env

4. Update .env with your configuration:

DATABASE_URL=postgres://user:password@localhost:5432/tamil_proofreading?sslmode=disable
PORT=8080
FRONTEND_URL=http://localhost:3000
JWT_SECRET=your-secret-key-change-in-production
OPENAI_API_KEY=your-openai-api-key
STRIPE_SECRET_KEY=your-stripe-secret-key
RAZORPAY_KEY_ID=your-razorpay-key-id
RAZORPAY_KEY_SECRET=your-razorpay-key-secret

5. Run the server:

go run cmd/server/main.go

The server will start on port 8080 and automatically migrate the database schema.

Frontend Setup

1. Navigate to the frontend directory:

cd frontend

2. Install dependencies:

npm install

3. Set up environment variables (copy .env.example to .env.local):

cp .env.example .env.local

4. Update .env.local:

NEXT_PUBLIC_API_URL=http://localhost:8080/api/v1
NEXT_PUBLIC_RAZORPAY_KEY_ID=your-razorpay-key-id
NEXT_PUBLIC_STRIPE_PUBLISHABLE_KEY=your-stripe-publishable-key

5. Run the development server:

npm run dev

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

API Documentation

Authentication

• POST /api/v1/auth/register - Register new user
• POST /api/v1/auth/login - Login user
• GET /api/v1/auth/me - Get current user (protected)

Submissions

• POST /api/v1/submit - Submit text for proofreading (protected)
• GET /api/v1/submissions - Get user submissions (protected)
• GET /api/v1/submissions/:id - Get submission by ID (protected)

Payments

• POST /api/v1/payments/create - Create payment (protected)
• POST /api/v1/payments/verify - Verify payment (protected)
• GET /api/v1/payments - Get user payments (protected)

Dashboard

• GET /api/v1/dashboard/stats - Get dashboard statistics (protected)
• GET /api/v1/usage - Get usage statistics (protected)

Admin

• GET /api/v1/admin/users - Get all users (admin)
• PUT /api/v1/admin/users/:id - Update user (admin)
• DELETE /api/v1/admin/users/:id - Delete user (admin)
• GET /api/v1/admin/payments - Get all payments (admin)
• GET /api/v1/admin/analytics - Get analytics (admin)
• GET /api/v1/admin/model-logs - Get model logs (admin)

Webhooks

• POST /api/v1/webhooks/stripe - Stripe webhook
• POST /api/v1/webhooks/razorpay - Razorpay webhook

Security Features

• JWT Authentication: Secure token-based authentication
• Rate Limiting: 100 requests per minute per IP
• Input Sanitization: XSS protection for user inputs
• CORS Protection: Configurable CORS middleware
• Password Hashing: bcrypt for password storage
• Role-Based Access Control: Writer, Reviewer, Admin roles

Pricing Model

• Pay-per-use: ₹10 per 500 words (Model A), ₹20 per 500 words (Model B)
• Subscription Plans:
  • Basic: ₹500/month - 10,000 words (Model A)
  • Pro: ₹1500/month - 50,000 words (Model B)
  • Enterprise: Custom pricing

Deployment

Backend

The backend can be deployed to:

• Railway
• Render
• Google Cloud Run
• AWS ECS
• Any platform supporting Go applications

Frontend

The frontend can be deployed to:

• Vercel (recommended)
• Netlify
• AWS Amplify
• Any platform supporting Next.js

License

This project is licensed under the MIT License.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

Support

For support, please open an issue on GitHub or contact the development team.