0560GitHub
Slop Filter: score and collapse low-quality social posts
A browser extension asks Jev narrow questions about writing patterns, then combines their probabilities to filter social feeds.
The public extension and System One client were inspected on October 1, 2026. Jev scores writing-pattern questions; deterministic weights produce the final score. Users set thresholds, reveal hidden posts, and label examples to fit their own weights. The extension supports multiple social platforms and provider choices. The scores are heuristic judgments, not proof that a post was written by AI; no independent detection-accuracy claim is made here.
adamnroman/slop-filterREADME ↗
slop-filter <img src="assets/icons/icon-128.png" alt="" width="40" height="40" align="top"></h1> The goal of this project is to collectively pursue an internet without having to sift through slop. It doesn't end at posts on X or comments on Reddit. It also means blocking bot accounts, YouTube and TikTok videos with AI-generated scripts, and whatever comes next. Today it is a Chrome and Firefox extension that hides AI-generated posts and comments on X, LinkedIn, and Reddit, and comments and videos on YouTube. A video is judged by its script, never by its title or thumbnail. ## What it does - Scores every post in your feed for how likely it is to be AI-written. - Collapses the ones over your threshold. One click shows a post again. - Animated mode lets you watch it work: a scan line runs over each post, then the post turns green and stays, or turns red and folds away. - Can be tuned to your own judgment. Mark posts as `AI` or `Human`, then fit the weights from your labels. - Blocks accounts. After three flagged posts from one account it offers to block it. A blocked account's posts are hidden on sight and never scored again. Hover any score for a Block button, and manage the list on the options page. ## How it does it 1. It reads the text of each post as it nears your screen. 2. It asks [TypeSafe Jev](https://docs.typesafe.ai) 30 narrow questions about how the text is written, not what it says. Human tells come first: an invented word, a joke, a stretched or repeated word, a typo. A confident one settles it as human. Then the AI tells. A person can be generic or repeat the post they answer, and that is not AI. The tells are in the cadence: the "it's not X, it's Y" pivot, three matching beats in a row, a tidy bow on the end, words that add importance and no informa