Chinese AI Text Humanizer
Humanizes Chinese AI-generated text and inspects hidden watermarks through a six-stage, zero-dependency offline pipeline.
Humanize Chinese provides a comprehensive six-stage pipeline designed to strip artificial phrasing, mechanical transitions, and cliché buzzwords from Chinese AI text without relying on external APIs. Combining native LLM writing guidelines with local Python CLI tools, it handles genre-aware vocabulary replacement, De-Clauding conversational habits, and watermark inspection for zero-width characters and homoglyphs. The skill preserves factual accuracy and rhythm across academic papers, social posts, and long-form writing while providing realistic scoring to prevent over-optimization and unnatural phrasing.
Características Principales
01Claude assistant tone reduction based on localized claudish-to-english translation rules
0229 GitHub stars
03Genre-aware offline phrase replacement supporting academic, social media, general, and novel styles
04Comprehensive watermark detection and cleaning for zero-width characters, bidirectional marks, and homoglyphs
05Six-stage humanization pipeline covering generation guidelines, targeted rewriting, vocabulary routing, and proofreading
06Statistical rhythm and sentence-length optimization without introducing grammatical errors
Casos de Uso
01Auditing and cleaning invisible metadata, zero-width spaces, and sampling watermarks from generated Chinese text
02Converting robotic AI marketing and social media copy into natural, engaging human-written Chinese
03Polishing academic papers and thesis abstracts to bypass false-positive AIGC detection while preserving formal tone