Humaniser

•
46
•

Refines AI-generated content by removing mechanical patterns and injecting natural, human-sounding rhythm and personality.

Humaniser is a specialized Claude Code skill designed to strip away the tell-tale markers of LLM-generated text and replace them with a genuine human voice. Based on extensive documentation from Wikipedia’s AI cleanup project, it identifies and fixes 28 distinct patterns—such as overused vocabulary like 'delve' or 'tapestry', formulaic structures, and excessive hedging—to transform sterile drafts into engaging, opinionated, and rhythmically varied prose. It is an essential tool for editors and developers who need their documentation, blog posts, or communications to feel authentic, persuasive, and distinctly non-algorithmic.

Características Principales

01Adapts and maintains the intended voice across formal, casual, or technical contexts
02Eliminates repetitive 'AI-isms' like inflated significance and superficial analyses
03Injects natural sentence variety and emotional resonance to give text a pulse
0446 GitHub stars
05Detects and fixes 28 specific AI writing patterns based on Wikipedia standards
06Removes chat artifacts, sycophantic tones, and knowledge-cutoff disclaimers

Casos de Uso

01Cleaning up marketing copy to remove cliché promotional language and vague attributions
02Polishing technical documentation to sound less robotic and more professional
03Converting LLM-generated drafts into blog posts that feel authentic and opinionated