Show HN: Detecting AI slop with regex and Stephen King (github.com)

A new tool called Tacheles has been introduced, aimed at detecting and correcting common writing flaws indicative of AI-generated text. It functions as an offline linter that identifies what the developer terms "AI slop," which includes unnecessary bloat and machine-like stylistic choices. Rather than generating new text, Tacheles flags specific lines in the input and provides actionable advice in the form of a cut list—allowing users to refine their writing and maintain their unique voice. The tool is based on established literary guidelines, drawing insights from renowned authors like Stephen King for English and Ильяхов and Нора Галь for Russian. The significance of Tacheles lies in its deterministic and user-driven approach to enhancing writing quality, distinguishing it from other AI detection tools that merely score texts. By providing clear reasons and solutions for each flagged item, it empowers writers to make informed edits. The tool is multilingual, supporting English, Russian, and experimental Hebrew, with future expansion plans for German and Spanish. As Tacheles operates entirely offline, it ensures privacy and reproducibility in results, offering a vital resource for writers looking to refine AI-assisted drafts without sacrificing their authenticity.