Show HN: API to detect abusive content in 27 languages
I am Vadim of Tisane Labs (http://tisane.ai). We build text analytics APIs.
My natural language processing journey started over a decade ago, when I built an NLP engine with an emphasis on scalability across languages. The engine served as a foundation of a startup I cofounded in 2010 (https://en.wikipedia.org/wiki/LinguaSys), which had Mark Cuban as an investor and was acquired in 2015.
Utilizing our expertise, we're trying to tackle one of the most painful issues with the Internet today: abusive content. Hate speech, sexual harassment, cyberbullying are difficult to detect. Today's solutions mostly look for obscenities; correlation with training datasets only works with limited demos as the patterns of abuse are too many. Hate speech is more than just ethnic slurs.
The API is a generic analytical engine, so it may be used for other purposes as well, but abusive content is our main focus.
We'd love to hear your feedback, ideas, wishes, and complaints.
4 comments
[ 3.7 ms ] story [ 24.0 ms ] threadWhat was the issue? Any error messages or you just didn't get the output you wanted?
I imagine you're interested in Norwegian?
EDIT: I forgot to mention that you don't need to code to test-drive. You can simply dump your JSON in the console (the "Try it" button: https://dev.tisane.ai/docs/services/5a3b6668a3511b11cc292655...).
The error I'm getting is:
using these parameters:I will need the entire client code though. Let's switch to email - the one you logged on with is OK, right?
EDIT: sent.