Slopsift

SlopSift catches the recognizable patterns that mark AI-generated writing before anyone else notices them. The tool scans Markdown, prose, HTML, and source code comments looking for grammatical structures and phrasing habits that show up frequently in content produced by language models. It's not claiming to prove who wrote something, and the creators make this limitation explicit right on the homepage. What it does is flag the stylistic tells that make readers suspicious or that editors have learned to spot. For writers polishing drafts, developers reviewing documentation, or editors screening submissions, SlopSift offers a way to spot and address these patterns before publishing rather than getting called out after the fact. The analysis runs locally on your machine using a lightweight dependency parser, so there's no sending content to external servers for evaluation and no data leaving your environment.

The detection approach relies on rule-based checks rather than machine learning models. Each rule has a name you can look up, making the process transparent rather than a black box verdict that offers no explanation. When SlopSift flags a passage, you can see exactly which pattern triggered it and why that pattern is considered characteristic of AI output. Maybe it's an overused transition phrase that language models reach for constantly, a particular sentence structure they favor, hedge words appearing with suspicious frequency, or opening patterns that recur across model outputs. This transparency lets you make informed decisions about what to revise rather than blindly accepting or ignoring a score you can't interpret. The deterministic nature also means running the same text twice produces the same results, unlike probabilistic AI detectors that sometimes flip their verdicts on identical content depending on random factors in their evaluation.

There are multiple ways to use SlopSift depending on your workflow and technical comfort. A browser-based editor on the website lets you paste text and get immediate feedback without installing anything or creating an account. You can experiment with it right away to see how your writing fares. For developers who want to integrate detection into their build process, content pipeline, or editorial tooling, there's an npm package available for direct installation. The entire codebase sits on GitHub under an open-source license, so you can inspect the rules in detail, understand exactly what patterns get flagged and why, suggest improvements based on your observations, or fork it for custom needs specific to your domain. Documentation walks through installation and configuration for those embedding it in larger systems where automated checking makes sense.

The target audience includes anyone producing or reviewing written content who cares about authenticity or wants to improve their style. Writers using AI assistance for drafts might run their work through SlopSift to identify sections that read too generically and need a personal touch or stronger voice. Documentation teams can screen technical writing before it ships, catching patterns that undermine reader trust in instructional content. Publishers and educators reviewing submissions get a first-pass filter for material that might warrant closer inspection or questions to the author. Content marketers can check copy before it goes live. Because the tool processes everything locally, sensitive or proprietary content never leaves your environment, making it safe to use with confidential material. There's no account to create, no data leaving your machine, no usage logging, and no ongoing subscription to maintain.

SlopSift explicitly avoids making claims about authorship determination, and this honest framing matters for how you use the results appropriately. Finding patterns associated with AI writing doesn't mean a human didn't write it since humans absolutely can and do use these same patterns. Absence of those patterns doesn't guarantee authenticity either since skilled users can prompt AI to avoid common tells. What you get is a practical editing tool that highlights stylistic characteristics worth reconsidering, not a forensic verdict you'd want to make accusations based on. This realistic positioning makes it more useful for its intended purpose of improving writing quality rather than policing authorship, which is a different and much harder problem. The focus stays on actionable feedback you can actually apply to make your text better rather than generating controversy you can't back up.

The tool fills a gap between ignoring AI writing patterns entirely and relying on opaque commercial detectors with questionable accuracy that produce scores without explanations. By keeping rules named and inspectable, processing local, and expectations realistic about what it can and cannot determine, SlopSift offers a lightweight way to check content without the drama that surrounds AI detection tools in general. Whether you're cleaning up your own drafts that started from AI prompts, screening incoming content from freelancers, or just curious about what patterns appear in your natural writing, having a transparent pass at common AI writing habits beats catching them after publication when readers call it out in comments or social media threads.