The discussion is about Filtering spam/troll accounts. haydenjames is looking for a way to disable email-only signup and only allow signups via GitHub, Twitter, or Discord to prevent a user from repeatedly logging in and posting spam content using VPNs.
supermathie suggested disabling local logins, but this may lock out existing members. haydenjames asked about login security or user moderation plugins that can help with this issue.
Alternative solutions were proposed, including using the anti-spam AI module by supermathie and sam, hiding email signup fields and replacing them with OAuth providers by NateDhaliwal, and using the Discourse Fingerprint plugin by Architect.
haydenjames considered trying the anti-spam AI module, and sam clarified that it is safe to use in production. haydenjames then confirmed the correct steps to install and use the plugin.
Is there a way to disable email-only signup? I would like to set up so that signup can only be via GitHub, Twitter or Discord.
I have a user who is using VPNs to repeatedly log in and troll or post 4h1t content.
With posts like these, lol:
Also open to alternative approaches, plugins or solutions. Thanks!
Thanks. Sounds like this may lock out at least some existing members? And login security or user moderation plugins available that can help with this?
I know people have had a lot of success using the anti-spam AI module?
What if you hid the email signup fields and replaced them with them pointing them to login with OAuth providers?
There is the fingerprinting option not sure if that would work for you:
If you’ve already banned their I.P. and e-mail vpn wouldn’t make any difference if they use the same e-mail, but guess they are using different e-mails and I.P.s then so ban isn’t working.
Don’t know what “4h1t” means, that is a strange post there Ammonia and Bleach are definitely not safe to be mixed that generates toxic gasses.
I would like to second this, it is free (using Gemini for example) and you could have custom instructions that block even if the troll becomes more sophisticated.
The trolls experience would be quite bad as well, cause they would get block right away and only you would see the spam in your queue which you can easily delete.
Ok I may give this a try. It’s not officially supported though. Safe to use in production?
Can you expand on the question here? I am not following, are you saying Gemini is not officially supported?
I think I misunderstood. Please ignore my ignorance. 
So to make use of this I need to install this plugin: Discourse AI
Then, follow this guide?: Discourse AI - Spam detection
Let me know if I understand correctly.
