OpenCensorAPI - My Final Year Project

This is my school project for final year in Software Engineering.

What is this?

OpenCensorAPI is a simple service made with PHP. It acts as a middle layer API to check Hebrew text for bad words. It connects to an AI service on Modal that uses the OpenCensor-Hebrew model. This API gives data to a separate frontend with user interface.

What it can do

• Check one Hebrew text for bad words.
• Check up to 256 texts at once.
• Uses simple HTTP calls with JSON answers.
• Works with requests from other websites (CORS).
• Easy to put on any PHP host.

How it works

This API service (middle layer):

• PHP code that handles requests from frontend.
• Manages web calls and sends to AI backend.
• Can go on any PHP server.

AI backend:

• See repo: https://github.com/LikoKIko/OpenCensor-Modal
• Uses FastAPI with fast computer help.
• Runs OpenCensor-Hebrew model from Hugging Face.
• Based on AlephBERT for Hebrew.
• 95% right, trained on over 3,000 Hebrew texts.
• Gives score from 0 (good) to 1 (bad).

How to install

1. Get Composer.
2. Run composer install for needed files.
3. Copy .env.example to .env and add:

   OPENCENSOR_URL=https://your-ai-endpoint.com
   

How to use the API

POST /predict

Check one text.

Ask:

{
  "text": "זה טקסט כלשהו :)"
}

Answer:

{
  "data": [0.05, "CLEAN"]
}

POST /batch

Check many texts (max 256).

Ask:

{
  "texts": ["זה טקסט כלשהו 1", "זה טקסט כלשהו 2"]
}

GET /

Info about API and ways.

Project files

├── api/
│   └── index.php     # Main API code
├── includes/
│   └── oc.php        # Class to talk to AI
├── composer.json     # Needed PHP things
├── vercel.json       # For Vercel setup
├── .env.example      # Sample settings
└── README.md         # This file

What you need

• GuzzleHttp: For web calls.
• vlucas/phpdotenv: For settings.

How to put it online

Works on any PHP host with:

• PHP 7.4 or more.
• Composer.

Steps:

1. Upload files to server.
2. Run composer install.
3. Set .env with AI link.
4. Point server to api/ folder.

Other related stuff

• AI Service: https://github.com/LikoKIko/OpenCensor-Modal
• Model: https://huggingface.co/LikoKIko/OpenCensor-Hebrew
• Demo: https://huggingface.co/spaces/LikoKIko/OpenCensor

License

This is a school project for final year in Software Engineering.

Give Credit: If you use this project or part of it in public, business, or school work, you must give credit to the maker. Include:

• Author: LikoKiko
• Project: OpenCensorAPI - Hebrew Bad Words Checker
• Repository: https://github.com/LikoKiko/OpenCensor-API

How to Use:

• Free for school and research.
• Free to change and share if you give credit.
• Okay for business if you give credit.
• No promise it works, no blame if problems.

Example Credit:

Based on OpenCensorAPI by LikoKiko
Original repository: https://github.com/LikoKiko/OpenCensor-API