In Progress · ai
Unfold AI
Browser security extension for AI-assisted phishing and malicious-site detection.
Proprietary license · Not open source · Not MIT
Unfold AI is a browser security extension designed to help users identify and prevent access to phishing and potentially malicious websites. The project combines browser-level protection with AI-driven analysis to address one of the most common cybersecurity risks faced by everyday internet users: being tricked into visiting a fraudulent website.
The extension is designed to work at the point where the threat occurs — the user's browser. Instead of relying entirely on users to recognise suspicious URLs, misleading pages or fake login websites themselves, Unfold AI can analyse a website and determine whether it presents characteristics associated with phishing.
The core objective is prevention rather than simply reporting an incident after the user has already interacted with a malicious site. By intervening when a potentially dangerous website is encountered, the extension can provide an additional layer of protection between the user and the phishing page.
The project is particularly relevant because phishing attacks increasingly use convincing copies of legitimate websites, making visual identification difficult even for technically knowledgeable users. Domain names, page content, branding, login forms and other characteristics can all be manipulated to make fraudulent websites appear legitimate.
The AI component provides an opportunity to analyse these signals and assist in making a risk assessment. This positions Unfold AI differently from a simple blocklist-based browser extension: its purpose is to apply intelligent analysis to determine whether a website should be considered suspicious.
The name was originally Unfurl AI and was subsequently changed to Unfold AI. The project represents applying AI to a practical cybersecurity problem rather than building an AI product purely for demonstration — combining browser-extension technology, cybersecurity and AI-assisted phishing detection into a consumer-focused security product.
Problem
Phishing sites increasingly clone legitimate brands so convincingly that users cannot reliably spot fraud by eye alone.
Solution
A browser extension that analyses sites at visit time with AI-assisted signals — prevention at the moment of risk, not after the fact.
Features
- In-browser phishing risk intervention
- AI-assisted site analysis
- Beyond simple blocklists
- Consumer-focused security UX
- Prevention-first design