unslop

Run the draft through the unslop rule set: scan for the listed patterns, rewrite each instance while preserving meaning and tone, then self-audit by asking what still makes the text obviously AI generated and fixing what remains.

Common tells include words like delve, showcase, and tapestry, phrases like "serves as" instead of "is", superficial -ing phrases, vague attributions like "experts believe", em dash overuse, and chatbot closers like "I hope this helps!"

No. The process explicitly preserves meaning and matches the intended tone. It replaces fancy or vague wording with plain equivalents, cuts filler, and restructures sentences without altering the underlying claims.

Avoid applying them when a stylistic choice is deliberate, such as a bold lead-in that names an item and adds genuinely new detail, which the rules explicitly allow. Passive voice is also acceptable when the actor is unknown or irrelevant.

Yes. The rules target exactly the problems common in AI-drafted docs: abstract metaphor nouns like substrate and vector, feeling-based claims instead of mechanisms, and over-compressed fragments. It pushes text toward concrete facts and numbers.