--- START OF FILE goat_storytelling_agent/README.md ---

Fiction Fabricator -- GOAT Storytelling Agent Enhanced Edition

This project generates long-form fiction (short stories, novels) from a user-provided topic. It uses large language models (LLMs) and a structured, iterative writing process.

Key Improvements

• Simplified User Interaction: A text-based menu guides users.
• Streamlined LLM Interaction: Fewer LLM calls for efficiency.
• GLHF API: Access to powerful LLMs.
• Clearer Codebase: Improved readability and maintainability.
• Enhanced Error Handling: Robust handling and feedback.
• Automatic Scene Generation and Enhancement: Automated workflow for writing and expanding scenes.
• Title Generation: LLM generates a title based on the topic.
• Output Caching: Saves LLM responses to markdown files.
• Model Selection: User can choose from a list of LLMs.

Features

• Book Specification: Creates and refines story details.
• Plot Outline: Develops a three-act plot with chapters.
• Scene Breakdown: Divides chapters into scenes with specifications.
• Scene Writing: Generates scene text (dark, cynical, erotic).
• Scene Continuation: Expands existing scenes.

Getting Started

1. Prerequisites:

   • Python 3.9+
   • GLHF authentication token (GLHF_AUTH_TOKEN environment variable)

2. Installation:

   cd goat-storytelling-agent-enhanced
   python3 -m venv .venv
   source .venv/bin/activate
   pip install -r requirements.txt

3. Usage:

   python menu.py

   Follow the prompts to provide a topic and navigate the menu.

Configuration

config.py:

• MODEL_NAME: LLM name on GLHF (keep hf: prefix for Hugging Face models).
• TIMEOUT: Timeout for LLM calls (seconds).
• RETRY_WAIT: Wait time before retrying (seconds).

Contributing

Contributions are welcome! Open an issue or submit a pull request.

License

MIT License. Original GOAT Storytelling Agent by Thomas Leon Highbaugh. This enhanced edition is maintained by Thomas Leon Highbaugh.

Acknowledgements

• Thanks to GLHF.