Hello, I’m currently being flooded with these posts. I’ve closed registration because I couldn’t find any other setting to stop it.
Do you have any tips for me?
AI-generated summary
Tealk is experiencing a spam flood and has closed registration to mitigate it, seeking advice. awesomerobot recommends using Discourse’s spam prevention guide, suggesting AI spam detection or watched words for consistent spam phrases. Tealk expresses concern about LLM costs, but RGJ clarifies they’re not expensive and shares non-LLM tips from the guide. Canapin links to a thread on using watched words to block recent spam attacks, noting other forums—including theirs—have faced similar issues. NateDhaliwal suggests free models on OpenRouter, like Gemini. merefield emphasizes that AI isn’t necessary—manually reviewing first posts, banning bad users and their IPs can resolve the issue quickly, as they did on their own forum. piffy agrees, calling manual review effective if time permits. The thread collectively offers cost-effective, non-AI solutions alongside AI options for spam control.
We’ve got a guide that explains some options here:
AI spam detection would definitely catch posts like that, but you may also have luck with the watched words feature if there are some consistent phrases in the spam posts.
Hmm, but then I need an LLM provider, and they’re all expensive.
They’re not really expensive.
And I’ve shared two tips in that topic that don’t need an LLM.
Hi, yes, this can be helpful:
Other Discourse forums have been the targets of this attack (mine included).
Or you can look at free models on OpenRouter, like Gemini.
You don’t need AI to solve this btw. Just initially a bit of work to review all first posts then ban the bad users and their IPs.
My forum had a very similar attack very recently and after doing that the situation was under control within half a day.
this is the way if you have the bandwidth to do so