Developer and security researcher David Dale has released Glassbox, a tool designed to demonstrate how visible a user is to modern browser fingerprinting systems. The service performs more than 30 checks that advertising trackers, anti-fraud platforms, and other entities commonly use to identify visitors without relying on cookies.
Among the attributes collected are Canvas rendering, WebGL capabilities, installed fonts, WebAssembly support, available APIs, authentication status on third-party sites, and audio processing behavior. Nearly all computations occur locally inside the browser, with the only external request made to a public API for geolocation data.
Users receive raw test results along with a calculated identifiability score. The developer stresses that the score is derived from a mathematical model rather than comparison against an actual database of visitors. Glassbox sums the uniqueness of individual attributes, accounts for built-in browser protections, and caps the result at approximately 33 bits—enough in theory to distinguish one individual among the entire world population.
In tests conducted by The Register, a standard Chrome browser scored 99 percent identifiability and was presumed unique among 7.6 billion browsers. Tor Browser scored 56 percent while Firefox reached 89 percent. These percentages should not be treated as definitive because the tool lacks server-side statistics on how rare each fingerprint actually is in the wild.
Dale notes that the most effective way to reduce visibility is to use a browser that blends into a large group of identical users. Overly customized or hardened configurations can backfire, as a unique combination of disabled APIs may itself become a strong identifier. He also recommends combining VPN or Tor with measures to block WebRTC leaks.
