I recently published a bot detection library based on browser fingerprinting: FP-Scanner. In order not to be detected simply because of their user agent, bots tend to modify their user agent to pretend to be a legitimate browser. For example, crawlers based on Chrome headless may modify their user agent from Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) HeadlessChrome/69.0.3071.115 Safari/537.36 to Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.77 Safari/537.36 to look like a normal Chrome.

Nevertheless, as I explained in two previous posts (post 1 and post 2), Chrome and Chrome headless have differences in their fingerprint that can be exploited to detect them. FP-Scanner’s goal is to verify if some fingerprint attributes related to known bots such as Chrome headless or Phantom JS are present in the fingerprint of a browser.

You can see a demo of FP-Collect, the library used to collect fingerprints, and FP-Scanner on this page. When you visit the page, the website automatically collects the fingerprint of your browser, which helps me improve the library by fixing tests that have false positives and false negatives.

For programmers that want a simple page that can be used by a crawler to test its fingerprint, you may want to visit this page. When the page is loaded, you can run the following code to obtain both the value of the fingerprint as well as the output of FP-Scanner.

const fp = JSON.parse(document.getElementById('fp').innerText);
const scanner = JSON.parse(document.getElementById('scanner').innerText);

// It gives the following output:
CHR_BATTERY: {name: "CHR_BATTERY", consistent: 3, data: {…}}
CHR_DEBUG_TOOLS: {name: "CHR_DEBUG_TOOLS", consistent: 3, data: {…}}
CHR_MEMORY: {name: "CHR_MEMORY", consistent: 3, data: {…}}
HEADCHR_CHROME_OBJ: {name: "HEADCHR_CHROME_OBJ", consistent: 3, data: {…}}
HEADCHR_IFRAME: {name: "HEADCHR_IFRAME", consistent: 3, data: {…}}
HEADCHR_PERMISSIONS: {name: "HEADCHR_PERMISSIONS", consistent: 3, data: {…}}
HEADCHR_PLUGINS: {name: "HEADCHR_PLUGINS", consistent: 3, data: {…}}
HEADCHR_UA: {name: "HEADCHR_UA", consistent: 3, data: {…}}
MQ_SCREEN: {name: "MQ_SCREEN", consistent: 1, data: {…}}
PHANTOM_ETSL: {name: "PHANTOM_ETSL", consistent: 3, data: {…}}
PHANTOM_LANGUAGE: {name: "PHANTOM_LANGUAGE", consistent: 3, data: {…}}
PHANTOM_OVERFLOW: {name: "PHANTOM_OVERFLOW", consistent: 3, data: {…}}
PHANTOM_PROPERTIES: {name: "PHANTOM_PROPERTIES", consistent: 3, data: {…}}
PHANTOM_UA: {name: "PHANTOM_UA", consistent: 3, data: {…}}
PHANTOM_WEBSOCKET: {name: "PHANTOM_WEBSOCKET", consistent: 3, data: {…}}
PHANTOM_WINDOW_HEIGHT: {name: "PHANTOM_WINDOW_HEIGHT", consistent: 3, data: {…}}
SELENIUM_DRIVER: {name: "SELENIUM_DRIVER", consistent: 3, data: {…}}
SEQUENTUM: {name: "SEQUENTUM", consistent: 3, data: {…}}
TRANSPARENT_PIXEL: {name: "TRANSPARENT_PIXEL", consistent: 3, data: {…}}
VIDEO_CODECS: {name: "VIDEO_CODECS", consistent: 3, data: {…}}
WEBDRIVER: {name: "WEBDRIVER", consistent: 3, data: {…}}

Feel free to contact me if you have questions or remarks about FP-Scanner or bot detection in general.

Antoine Vastel

Head of research at Datadome.