All filters are inherently problematic since the filters’ AI can’t tell whether the words in question are respectable or not. For example, if you target the word “naked” you block out halachic discussions about what one wears in a mikva, married women’s hair, and how much clothes you need to wear when davening.
The best solution is to remember that Ha-Shem is monitoring at all times (not to mention your boss, the NSA, perhaps the Mossad, and that any clever teenager can hack your computer to see what you’ve been looking at).