AiLens

AiLens is a full-stack web application designed to analyze text and scanned document images to determine whether content was generated by AI. The application features real-time AI probability scoring, supporting raw text input as well as image-based text extraction using optical character recognition (OCR).

---

Project Overview

• Frontend: Next.js 16, React 19, TypeScript, Tailwind CSS v4, and Tesseract.js for OCR.
• Backend: FastAPI with PyTorch and Hugging Face Transformers for inference.
• Model: desklib/ai-text-detector-v1.01 sequence classification model.

---

Frontend Setup

Dependencies

• Next.js (16.3.0): React framework for server rendering and static site generation
• React / React DOM (19.2.8): UI component library
• Axios (^1.19.0): HTTP client for backend API interaction
• Tesseract.js (^7.0.0): Optical character recognition engine for extracting text from images
• Tailwind CSS (^4) & PostCSS: Utility-first CSS styling
• TypeScript (^5): Static type checking

Installation and Usage

1. Navigate to the frontend directory:

   cd frontend

2. Install dependencies:

   npm install

3. Run the development server:

   npm run dev

4. Access the application at http://localhost:3000.

---

Backend Setup

Environment and Version Compatibility

Important Compatibility Note: The underlying model architecture is incompatible with newer versions of Hugging Face transformers.

• Python Version: Requires Python 3.12 (>=3.12).
• Transformers Version: Strictly pinned to transformers==4.44.2.

Dependencies

• transformers==4.44.2
• torch>=2.13.0
• fastapi>=0.141.1
• uvicorn>=0.52.1
• accelerate>=1.14.0
• pydantic>=2.13.4

Installation and Usage

Method 1: Using uv (Recommended)

1. Navigate to the backend directory:

   cd backend

2. Synchronize dependencies and set up the virtual environment:

   uv sync

3. Start the FastAPI backend server:

   uv run uvicorn main:app --reload

Method 2: Using Standard venv and pip

1. Navigate to the backend directory:

   cd backend

2. Create a virtual environment using Python 3.12:

   python3.12 -m venv .venv

3. Activate the virtual environment:

   • Windows (PowerShell):

     .venv\Scripts\Activate.ps1

   • Linux / macOS:

     source .venv/bin/activate

4. Install required packages:

   pip install "transformers==4.44.2" "torch>=2.13.0" "fastapi>=0.141.1" "uvicorn>=0.52.1" "accelerate>=1.14.0" "pydantic>=2.13.4"

5. Start the backend server:

   uvicorn main:app --reload

The backend server will be available at http://localhost:8000.

---

Model Credits

This application utilizes the open-weight AI text detection model:

• Model URL: https://huggingface.co/desklib/ai-text-detector-v1.01
• Model Name: desklib/ai-text-detector-v1.01
• Creator: Desklib

Credit belongs to Desklib for providing and publishing this AI text detection model.