Anonymous developer under the pseudonym Aaron B. presented the program Nepenthes – a tool created to neutralize web scanners that collect data for teaching artificial intelligence. Named after predatory nopentes plants, this tool forms an endless cycle of pages with links, closing the scanners in a trap and forcing them to waste resources.
The principle of operation of Nepenthes is simple: the program generates links that constantly return the bot back to the created labyrinth. According to Aaron B., most web scanners mechanically upload pages and follow the specified links without analyzing their contents. Thus, Nepenthes effectively locks them in endless wandering, depriving access to real content.
The program website posted a warning that this is “intentionally malicious code”, which should only be used consciously. Nepenthes can be used to protect the site content from unwanted scanning or to attack scanners, forcing them to spend computing power. “This is a way to complicate access to real content and put unnecessary resources to false data,” the developer notes.
Site owners have long come across difficulties in the fight against automatic scanners ignoring Robots.txt rules or in many ways to circumvent restrictions. An example is cases when Openai and Anthropic bots in one day made millions of requests on separate sites, such as fixit. These cases have become one of the reasons for the release of nepenthes.
The creator of the tool admits that his project is a form of digital protest. “The Internet has turned into a profit made tool controlled by large corporations, and the time has come to repulse,” Aaron B. According to him, the program is also a kind of art that expresses anger and disappointment from a modern digital environment.
Some argue that such traps are easy to get around, but millions of lines of Nepenthes logs indicate the opposite. Even the largest players, such as Google, were unable to effectively cope with this protective system.