Humanizer is a professional-grade editor that detects and removes robotic AI writing patterns to produce natural, human-sounding content.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ai-humanizer
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install ai-humanizer using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Humanizer?
Humanizer is a specialized writing editor designed to identify and eliminate the recognizable hallmarks of AI-generated content. By leveraging a comprehensive suite of 24 pattern detectors and over 500 specific AI vocabulary terms, it ensures your text resonates with an authentic human voice rather than the sterile output of a language model. This tool is a vital addition to your Openclaw Skills library, enabling the production of high-quality content that bypasses common detection markers while maintaining structural integrity.
Beyond simple word replacement, the skill employs statistical analysis to evaluate text burstiness and sentence variety. By focusing on nuances like type-token ratios and sentence length variation, it transforms mechanical drafts into nuanced narratives. Whether you are refining a technical document or a creative blog post, this skill provides the tools necessary to ensure your writing feels personal, specific, and genuinely human.
Humanizer Use Cases
- Rebranding AI-generated drafts into professional blog posts and articles.
- Scoring existing content for AI detection risk before final publication.
- Cleaning up chatbot artifacts and sycophantic phrases in automated reports.
- Enhancing content for digital marketing campaigns to ensure high engagement through natural language.
- Auditing technical documentation to remove formulaic structures and significance inflation.
How Humanizer Works
- The skill scans the input text against 24 heuristic patterns, including significance inflation, promotional language, and formulaic challenges.
- It cross-references content with a three-tier vocabulary database of over 500 words and phrases common to LLM outputs.
- Statistical indicators are calculated to determine burstiness, vocabulary diversity (TTR), and sentence length uniformity.
- Problematic sections are flagged or automatically rewritten using natural, human-like alternatives that preserve core meaning.
- The tool generates a final output, ranging from a simple humanized text to a detailed markdown report for deeper manual editing.
Humanizer Setup
Install the skill and use the CLI to begin analyzing your text immediately:
# Score text to see how AI-like it is (0-100)
echo "Your text here" | node src/cli.js score
# Generate a full markdown report for an article
node src/cli.js report article.txt > report.md
# Automatically fix and humanize a file
node src/cli.js humanize --autofix -f article.txt
Humanizer Data Schema & Taxonomy
Humanizer organizes its analysis through a rigorous taxonomy of patterns and statistical metrics:
| Category | Metrics and Patterns |
|---|---|
| Content | Significance inflation, vague attributions, formulaic challenges |
| Language | 500+ AI vocabulary terms, copula avoidance, negative parallelisms |
| Style | Em dash/boldface overuse, Title Case headings, emoji frequency |
| Statistics | Burstiness (0.5-1.0), Type-token ratio (0.5-0.7), Trigram repetition (<0.05) |
| Communication | Chatbot artifacts, sycophantic tone, cutoff disclaimers |
Humanizer Advanced Features
- Always-On Mode: Integration templates for Openclaw Skills (SOUL.md) to ensure agents write naturally by default.
- Multi-Tier Vocabulary Filtering: Differentiates between dead giveaways like delve or tapestry and suspicious density terms like paradigm.
- Statistical Tell Detection: Identifies metronomic sentence rhythms and low sentence length CoV (Coefficient of Variation).
- Programmatic JSON Output: Allows developers to integrate AI detection and humanization directly into automated content pipelines.
SKILL.md
Loading
Related Openclaw Skills
Platform API Connector
A comprehensive automation tool for navigating developer portals and managing OAuth credentials for major social media and content platforms.
Domain Email Forwarding
A technical guide for routing custom domain emails to existing inboxes using Openclaw Skills to facilitate account recovery and catch-all addresses.
Yr.no Weather Forecast
A Python-based integration for fetching precise meteorological data and forecasts from MET Norway via coordinate-based lookups.
Training Delivery Manager
An all-in-one automation skill for managing the end-to-end lifecycle of corporate training projects, from legal contracts to promotional content.
API Security Best Practices
A comprehensive framework for implementing secure API design patterns, authentication, and protection against common vulnerabilities.
Secrets Management
A comprehensive skill for securing API keys, database credentials, and TLS certificates across multi-cloud environments and CI/CD pipelines.
Featured*
