Last updated: May 8, 2026
These terms govern the use of the aisafemod application ("the App") on Reddit subreddits.
By installing the App on a subreddit, the moderator who performs the installation accepts these terms on behalf of the subreddit. By posting comments on a subreddit where the App is installed, users implicitly accept these terms with respect to the moderation activity described herein.
The App provides automated comment moderation. It uses OpenAI's omni-moderation API to classify comments into 13 content categories (sexual, harassment, hate, violence, self-harm, illicit, plus subcategories) and removes comments whose category scores exceed thresholds configured by the moderator who installed the App.
Automated moderation is inherently imperfect. The App may produce false positives (legitimate comments classified as violations) and false negatives (violating comments not flagged). The OpenAI Moderation model is best calibrated for English; non-English content may produce different score distributions. The App is provided "as is" without warranty of accuracy, completeness, or fitness for any particular purpose.
Moderators of subreddits where the App is installed are responsible for:
The App is a tool that augments moderator decision-making. It does not replace human judgment.
The App's actions are limited to:
The App does not: