humanize-ai-text

What this skill does

# humanize-ai-text

## Purpose
This skill humanizes AI-generated text by reducing detectable patterns, enhancing natural phrasing, and adjusting tone to make it indistinguishable from human writing. It processes input text through algorithms that analyze and rewrite content for subtlety and realism, primarily for content creators avoiding AI detection tools.

## When to Use
Use this skill when you need to refine AI-produced text for blogs, articles, or social media to evade detection by tools like Originality.ai or GPTZero. Apply it in scenarios involving content marketing, academic writing, or any context where text must appear authentically human, such as rewriting product descriptions or email drafts generated by AI.

## Key Capabilities
- Reduces AI-specific patterns like repetitive structures or unnatural vocabulary by applying natural language processing techniques.
- Adjusts tone (e.g., formal to casual) based on user-specified parameters, using predefined models for styles like conversational or professional.
- Handles text lengths up to 10,000 characters, preserving original meaning while enhancing readability and engagement.
- Integrates sentiment analysis to ensure the output aligns with desired emotional tones, such as positive or neutral.

## Usage Patterns
Always provide input text via file, string, or API payload; specify output format and tone adjustments explicitly. For CLI, pipe input directly; for API, use JSON requests. Process text in batches for efficiency, and review outputs for context-specific tweaks. Avoid using on highly creative or poetic text, as it may alter nuances unintentionally.

## Common Commands/API
Use the OpenClaw CLI for quick operations or the REST API for programmatic access. Authentication requires setting the environment variable `$OPENCLAW_API_KEY` before running commands.

- **CLI Command Example**:  
  `openclaw humanize --input "The quick brown fox jumps over the lazy dog." --tone casual --output output.txt`  
  This rewrites the input with a casual tone and saves to a file.

- **API Endpoint Example**:  
  POST to `https://api.openclaw.ai/v1/humanize` with JSON body:  
  `{ "text": "AI-generated content here.", "tone": "formal" }`  
  Response: JSON object with "humanized_text" key.

- **Code Snippet (Python)**:  
  ```python
  import requests
  headers = {'Authorization': f'Bearer {os.environ.get("OPENCLAW_API_KEY")}'}
  response = requests.post('https://api.openclaw.ai/v1/humanize', json={'text': 'Sample AI text.'}, headers=headers)
  print(response.json()['humanized_text'])
  ```

- **Config Format**:  
  Use a YAML config file for repeated tasks:  
  ```
  input_file: input.txt
  tone: neutral
  output_file: output.txt
  ```  
  Invoke with: `openclaw humanize --config config.yaml`

## Integration Notes
Integrate by setting `$OPENCLAW_API_KEY` in your environment or passing it via headers in API calls. For scripts, import the OpenClaw SDK and handle asynchronous requests for large texts. Ensure compatibility with Python 3.8+ or Node.js 14+; add error logging for API responses. If using in a workflow, chain with tools like text generators by piping outputs, e.g., via shell scripts.

## Error Handling
Check for common errors like invalid API keys (HTTP 401) by verifying `$OPENCLAW_API_KEY` is set and not expired. Handle input errors (e.g., empty text) with try-except blocks in code, returning user-friendly messages. For CLI, parse flags carefully to avoid "Invalid argument" errors; use `--help` for validation. If rate limits are hit (HTTP 429), implement retries with exponential backoff, e.g., in Python:  
```python
import time
time.sleep(5)  # Wait before retrying
```

## Graph Relationships
- Related to: "ai-writing" skill in the community cluster via shared tags ["ai", "writing"].
- Connected to: "text-analysis" skill for preprocessing text inputs.
- Linked by: "humanize" tag to "natural-language-tools" cluster for broader NLP capabilities.