mstfknn/phishing-fasttext repository overview

⁠🛡️ Phishing Detection Model (FastText)

A lightweight FastText model that classifies domain names as phishing or clean. Runs as a simple REST API inside Docker.

⁠🚀 Quick Start

Pull and run the container:

docker run -p 8080:8080 mstfknn/phishing-fasttext:latest

⁠🔎 API Usage

Predict a domain (POST /predict):

curl -X POST http://localhost:8080/predict \
  -H "Content-Type: application/json" \
  -d '{"domain": "carreeffoursa.site"}'

Example response:

{
  "domain": "carreeffoursa.site",
  "label": "__label__phishing",
  "probability": 0.9734
}

⁠📦 Notes

• This image exposes port 8080 for the HTTP API.
• The container includes the FastText model file phishing_model.bin at /app/phishing_model.bin by default. You can override the model path by setting the MODEL_PATH environment variable.

Example (mount custom model):

docker run -p 8080:8080 -v /local/path/phishing_model.bin:/app/phishing_model.bin mstfknn/phishing-fasttext:latest

⁠📊 Model Info

• Framework: FastText
• Labels: __label__phishing, __label__clean
• Training data: mstfknn/phishing-domain-list-2m-plus⁠

⁠📄 License

MIT License © mstfknn⁠

⁠🔗 Links

• 🤗 Hugging Face Model⁠
• 💻 GitHub Repository⁠