unslop

Scan the text for known AI tells like puffery, vague attributions, em dash overuse, and chatbot phrases, then rewrite each instance with plain words and concrete facts. Finish with a self-audit asking what still makes the text obviously AI generated.

Common tells include words like delve, pivotal, and tapestry, em dash overuse, rule-of-three structures, boldface on every noun, and phrases like "I hope this helps". Formulaic conclusions and excessive hedging are also strong indicators.

Yes, it applies to any prose including technical docs. It replaces abstract jargon like substrate, vector, and leverage with concrete words, converts passive voice to active, and cuts sentences that could appear unchanged in any project's documentation.

No, the process preserves meaning and matches the intended tone while rewriting. The goal is to state the same facts directly, replacing vague or promotional language with specific, neutral descriptions.

Removing patterns is only half the job; sterile voiceless writing is equally obvious. Add opinions, vary sentence rhythm, use first person where it fits, and include specific details instead of neutral summaries.