AI / ML Project

Phishing URL Detection

Paste a link and get a verdict. The URL's structure is checked in your browser, then a machine learning model and live domain checks decide if it looks like phishing.

  • Python
  • Scikit-learn
  • Logistic Regression
  • CountVectorizer
  • DNS / RDAP
  • Flask
enter url Nothing is opened or visited
Try an example
How to read the result
  • Phishing score: how much the link's text resembles known phishing links. Under 40% looks safe, 40 to 70% is suspicious, over 70% is likely phishing.
  • Format checks: quick tests of the link's shape, done in your browser.
  • Domain checks: whether the name exists and how old it is. The website itself is never opened.

prashant@linux:~$ model.predict(url)

Verdict --

Enter a URL and press Analyze.

Phishing score (the model's estimate from the link's text)
--
Looks safeSuspiciousLikely phishing
URL format checks
    Domain checks

      Educational demo, not a security product. A "safe" result never guarantees a link is safe.

      Dataset

      Write your dataset source, number of URLs and the phishing / legitimate split.

      Model

      Write which model you used and which URL features it learns from.

      Accuracy

      Write your test accuracy, precision and recall here.

      Source Code

      View on GitHub