Chat moderation pilot

As referenced in this article, Twitch will be piloting new tools to help reduce negative chat interactions. If you’re interested in being a part of this pilot program, please fill out the Google Form below to register your email for announcements. Otherwise, please feel free to leave feedback about what you’d like to see in these tools.

Fill out the Google form here.

1 Like

Couldn’t you just get a bot like hnlbot and setup regular expressions?

1 Like

Regular expressions are great, but not the ultimate solution to the problem. It also requires a 3rd party application to do it while an official solution is desired.

So its gonna be some really big server side bad word database?

I hope when there is a solution it will be toggleable like the bad words filter already in the settings or included in that.

I understand that twitch wants a official solution. But any solution other than regex won’t be as good as third parties have to offer now such as mentioned by cheesasian.

I doubt it. After all, that’s pretty much what the current word filter in the channel settings is.

Regex can only match language, not meaning. There will always be another way to convey the same message in a way that doesn’t match the regex meant to filter it.

I’m interested to see what it’s going to be like, but I highly doubt it’ll be regex due to it being difficult to use.

I disagree, first party tends to have a large advantage in combatting spam, because first party can fingerprint users with a wide variety of methods ranging from IP, login information, user behavior, registration dates, etc.

You can use word filters to create triggers on users, and moderators/staff/admins to create rules about what is bad behavior and issue timeouts/bans automatically.

Third parties don’t have the fingerprinting data, so they can only use a subset of available information to create rules to catch bad behavior.

As a moderator for several large CSGO streams in particular, it never ceases to amaze (dismay?) how people get around filters to spam referral links what with the umpteen betting sites that recently sprang up.

I’m really looking forward to what Twitch can deliver.


regular expressions should be perfect for this matter tho. And twitch probably just has people feeding the system with new patterns every day as well.

Yes alot to discuss here over the last 3 years+
mainly to do with psyche and hivemind mechanics besides usual coding and “on the fly” spots <3

Looking forward to any debates or brainstorming

I think the obvious solution is to assign permissions to trusted members of the community who will contribute to a large list of banned phrases. I plan on making some kind of chat logger that’ll log commonly used emotes and phrases that people use in racist messages. Hopefully that’ll help those people look through chat logs easier.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.