I have built an AutoMod bot for my instance, lemmy.basedcount.com. The bot covers the following features:
- Automated removal
- of posts, based on their title, content or link
- of comments, based on their content
- configurable with either regular expressions or substrings
- User whitelisting and exceptions for moderators to selectively lift some or all of the aforementioned rules for certain users.
- Mention based pinning and locking of a post, through commands exclusively available to the mod team
- Discord notifications for new registration applications through a webhook. [only for admins]
Naturally, the bot is completely open source. I have also written a rather comprehensive (albeit long-winded) documentation and some examples.
This project is mainly targeted towards admins of small instances, however anyone can spin up their own AutoMod instance for their favourite community (provided they are a moderator there).
The automoderator is also available as a Docker image, for ease of installation.
Feel free to suggest any additional features that you might want to see added to this bot.
Like anything, it’s how it’s implemented.
New users can find posting rules complicated, and getting constant automod rejections along with shadowbanning new users makes the whole thing tiresome.
Fuck that. I had that suggested as a feature and downright refused to implement it. It will never be a thing in my bot. This is entirely about reviewing content, not people.
There’s no way to discriminate one based on their account age, instance, what communities they follow or other bullshit systems such as Reddit’s karma.
Cool. :)
Although account-based filtering can be useful for managing obvious spam (e.g., a brand new account posting links).
That is true, but it’s a slippery slope I would rather not walk anywhere near. In my humble opinion, account based AutoMod restrictions were the first small steps for the enshittification of Reddit and I wouldn’t want to be responsible for such a thing happening to Lemmy.