Yeah unfortunately the neural network right now only works for “understandable” insults, to filter things like “ez” you would need a script that converts slangs into actual words
1 Like
Umm, no, but I am using a neural network which looks very promising to games, and I believe roblox should use something like this to aid in reports, for example
Hm weird, it seems exactly like this
Likely due to websites being tagged with “.gg” at the end.
It is fun putting in things mean people said to me in the past to see how toxic they are and indeed they’re very toxic
1 Like
Woah, didn’t realize someone else was using Perspective API, that guy didn’t tell anyone that he was using perspective. I am not using his work. Im using HTTP to request to a repl.it (free hosting) API made by me that uses my own perspective API Key. Thanks for telling me there was something similiar though.
And as already said by another member in his post, his module redirects directly to the API. All he is doing is providing a free API Key to others, while giving himself credit and not wanting others to use his code.
1 Like
Thanks for saying, going to delete the post.
It looks similiar because both of us use the same neural network API and both picked out categories in the message detection that wouldn’t bee inappropriate or explicit for roblox.
1 Like
UPDATE
Thanks to some suggestion, I decided to add threat.
Please note that messages are not filtered as no one sees them apart from yourself.
Is thing wrong to do, should I change it? Hit me up in the replies!
Wow that is really cool and interesting!