PhishShield: A Multi-Modal AI-Based Anti-Phishing Detection Platform

Phishing is a trick in which a fake website pretends to be a real one, such as a bank, to steal a person’s password or money. Many existing tools check only one thing, such as whether a link appears on a blacklist. A new fake site that is not yet listed can pass such a check. This paper proposes PhishShield, a platform that studies a website through five separate channels: the URL text, the HTML code of the page, the SSL certificate, the domain registration record (WHOIS), and a screenshot of the page. Each channel gives its own risk score between 0 and 1. A weighted sum then combines these scores into one final score. If the final score crosses a set limit, the user sees a warning and the event is saved in a log. The system is built as small separate services using FastAPI, so that each part can be changed without touching the others. This is a design paper. Testing is planned and is described in the Future Work section.