Without researcher input on the efficacy of Noisy, confidence might be low. As it stands, its design - fetching a publicly known set of webpages - makes its efficacy questionable, since website fingerprinting techniques can be particularly effective against a small, known subset of websites and pages. The automated nature of Noisy’s traffic generation also makes its patterns more easily distinguishable from natural user-generated traffic. At the time, the main advice from The Tor Project was that introducing different types of traffic beyond simple website browsing could create a more effective layer of noise? Since Noisy focuses solely on webpage traffic, its overall benefit might be limited.
Padding might be a better solution: