Discourse AI is phasing out its standalone NSFW detection module in favor of the more flexible and powerful AI Triage system powered by LLMs. Saif explains that LLMs offer superior customization, performance, and community-specific control compared to the old NSFW module. The transition includes guides for setting up NSFW, toxicity, and spam detection using AI Triage. Although initially planned for future deprecation, the NSFW module has now been officially removed, including all related site settings and features. Business and Enterprise users can enable Discourse-hosted LLMs for AI Triage at no extra cost via the admin panel. Saif confirms that AI Triage now supports image detection via vision-powered LLMs, making it fully capable of replacing the old NSFW functionality. Site settings for enabling Toxicity and NSFW have been hidden — existing enabled instances continue to function, but users can no longer re-enable them if turned off. Saif also notes that AI Triage no longer requires the %%POST%% placeholder in prompts, thanks to improvements by @Roman_Rizzi. Users are strongly encouraged to migrate to AI Triage using the provided guides, such as Setting up NSFW detection in your community.
This announcement should be considered very early, until we are ready to decommission you can continue to use NSFW. When we do it, we will be decommissioning the module, and removing all code from the Discourse AI plugin and associated services from our servers. The deprecation will only happen after AI Triage is ready to handle images which will be done soon.
Update: NSFW module has now been officially removed from Discourse, this includes all related site settings and features. We are now urging users to transition to using Discourse AI - AI triage and follow the guides listed above.
Business and Enterprise customers will see the following under What's New in the admin settings on their sites, allowing them to enable Discourse hosted LLMs to power AI triage at no extra cost.
As a heads-up, we are now hiding site settings for enabling/disabling Toxicity and NSFW. This is our ongoing effort as we continue to deprecate the features.
If you have these features turned on, it will still operate as normal. We have NOT fully deprecated the features yet.
If you have it turned off and would like to turn it on, you will now be unable to do so.
Update: NSFW module has now been officially removed from Discourse, this includes all related site settings and features. We are now urging users to transition to using Discourse AI - AI triage and follow the guides listed above.