How to track messages containing offensive nature, containing elements of violence, etc.?

Question

Python

Hello.
There is a project - social. open discussion network. There was a problem with the need to modernize these same discussions. The question is - are there any ready-made methods for optimal detection of messages that are offensive, containing elements of violence, incitement to national hatred, pornography ... and, if so, which ones?
Also interested, if the message contains a link to a third-party resource containing prohibited materials, etc., is it possible to somehow fix these things?
Thank you.

Answer 1

At the expense of links: you can work drastically and ban them altogether, if this is a bad option for you, then create an intermediate page where you will warn that this site can harm your computer, consciousness, and in general there may be boobs.
If this does not suit you, then try this API: https://www.mywot.com/wiki/API .
In order to get rid of people who want to post 18+ photos, there is an API: sightengine.com , it is paid. There is a free analogue from Microsoft - https://dev.projectoxford.ai/docs/services/54ef139...