blader/humanizer
Humanizer
Humanizer is a text processing system designed to remove machine-generated patterns from writing to make it sound more natural and conversational. It functions as an auditor and rewriter that identifies robotic signatures, formulaic tropes, and mechanical formatting in machine output.
The project features a style-matching system that analyzes provided writing samples to replicate a user's specific sentence rhythms, vocabulary, and punctuation habits. This allows the tool to mirror a personal voice and apply a calibrated tone to the rewritten text.
The system covers a broad range of linguistic refinements, including the conversion of passive voice to active constructions and the replacement of high-frequency machine vocabulary. It also employs a multi-pass auditing process to eliminate structural tropes, such as rule-of-three lists and generic conclusions.
Features
- AI Writing and Content - Removes common machine patterns from writing to make the content sound more natural and conversational.
- Natural Language Rewriters - Provides a system to convert passive voice and mechanical formatting into natural, human-like prose.
- Active Voice Converters - Detects subjectless fragments and passive constructions to rewrite them into direct and active sentences.
- AI Content Auditors - Identifies and eliminates robotic language signatures, formulaic structures, and common machine tropes.
- AI Content Editing - Cleans up formulaic structures, passive voice, and repetitive formatting common in machine generated drafts.
- AI Pattern Detectors - Eliminates inflated symbolism, promotional language, and formulaic structures to make content sound human.
- AI Language Artifact Auditors - Employs a multi-pass review process to identify and remove robotic language patterns and artificial signatures.
- AI Writing Pattern Detectors - Identifies formulaic structural habits like rule-of-three lists and generic conclusions using predefined machine-writing patterns.
- Authorial Style Mimicry - Analyzes a provided writing sample to apply specific sentence rhythms and word choices to rewritten text.
- LLM Text Humanizers - Removes machine patterns from LLM output to make AI-generated text sound more natural and conversational.
- Natural Language Refinement - Replaces high frequency AI vocabulary and promotional tropes with conversational phrasing and varied sentence structures.
- AI Vocabulary Substitutors - Identifies and replaces high-frequency machine words and phrases with more natural and varied language.
- Structural Trope Eliminators - Removes formulaic sections, rule-of-three lists, and generic conclusions indicative of machine generation.
- Text Feature Extraction - Analyzes sentence rhythm, vocabulary, and punctuation from user samples to create a quantitative profile for text replication.
- Text Refinement Passes - Iteratively scans generated text through different filters to identify and remove robotic signatures.
- Voice Replication - Analyzes sentence length, word choice, and punctuation habits to replicate a user's specific writing style.
- Writing Style Calibrators - Analyzes a writing sample for rhythm and vocabulary to apply a specific personal tone to the output.
- Linguistic Transformation Pipelines - Sequentially converts passive voice to active and replaces high-frequency machine vocabulary with natural alternatives.
- Prosody and Rhythm Adjusters - Adjusts sentence lengths and cadence to mirror the specific prosody and flow of a human writing sample.
- Formatting Artifact Removals - Detects and strips mechanical formatting tells such as excessive bolding and repetitive heading styles.
- Agent Skills - Tool for identifying and removing AI-generated writing patterns.
- Media and Communication - Removes AI-generated markers to improve natural text flow.
