Description:

This project provides a suite of Python tools designed to transform text, making it sound more natural, engaging, and human-like. Utilizing the power of Large Language Models (LLMs) through the Langchain library and OpenAI's API, these tools can take formal, robotic, or AI-generated text and refine it to better suit various communication styles.

Key Features:

• Iterative Humanization: Applies multiple refinement passes to progressively enhance the text's human-like qualities.
• Style-Specific Adaptation: Tailors the output to different tones and styles, including conversational, professional, casual, and academic.
• Contextual Awareness: Can consider additional context to ensure the humanized text is relevant and appropriate.
• Fact Preservation: Includes mechanisms to verify that key factual information from the original text is retained in the humanized version.
• Formality Analysis: Assesses the formality level of text to guide the humanization process.
• Batch Processing: Efficiently humanizes multiple text documents at once, with cost and token tracking.
• Interactive Mode: Allows for step-by-step humanization with user guidance and the ability to choose refinement options or provide custom instructions.
• Customizable Pipelines: Supports the creation of unique humanization workflows with user-defined processing steps.
• OpenAI Model Integration: Leverages models like GPT-4 for high-quality text generation and refinement.

Core Functionality:

The project offers several approaches to text humanization:

1. A basic iterative humanizer (src/ai2human/humanize.py) that applies a general humanization prompt multiple times.
2. A more structured class-based humanizer (src/ai2human/humanize2.py) that uses a sequence of specific prompts for iterative improvement and includes cost tracking.
3. An advanced humanizer (src/ai2human/humanize-adv.py) with sophisticated features like style targeting, formality analysis, fact verification, batch operations, and an interactive refinement loop.

Potential Use Cases:

• Improving the readability and engagement of AI-generated content.
• Making chatbot responses more natural and less robotic.
• Softening formal documents or emails for a wider audience.
• Enhancing marketing copy or creative writing.
• Assisting in the editing process to achieve a specific tone or style.

Technology Stack:

• Python
• Langchain
• OpenAI API (gpt-3.5-turbo, gpt-4, gpt-4o, etc.)
• Dotenv for environment management

This project is ideal for developers, writers, and content creators looking to bridge the gap between machine-generated text and human expression.

This project is Rye-enabled for easy script execution. After installing dependencies and setting up your .env, you can run all main tools using short aliases:

• rye run humanize --help
• rye run humanize2 --help
• rye run humanize_adv --help
• rye run ai_detector --help
• rye run ai_detector_adv --help

These aliases are defined in the [tool.rye.scripts] section of pyproject.toml and map to the main scripts in src/ai2human/.

You can pass any CLI arguments as usual, for example:

rye run humanize2 --iterations 2 --input "Your text here"
rye run humanize_adv --style conversational --input "Some text"

With Rye:

rye run humanize --help
rye run humanize --iterations 2 --input "Your text here"

Direct Python (also works):

python src/ai2human/humanize.py --help

With Rye:

rye run humanize2 --help
rye run humanize2 --iterations 2 --input "Your text here"

Direct Python (also works):

python src/ai2human/humanize2.py --help

With Rye:

rye run humanize_adv --help
rye run humanize_adv --style conversational --input "Some text"

Direct Python (also works):

python src/ai2human/humanize-adv.py --help

This project is ideal for developers, writers, and content creators looking to bridge the gap between machine-generated text and human expression.