Infinite Bookshelf is a Streamlit application designed for rapidly generating book-length content from a single prompt, targeting users who need to quickly create informational or educational material. It leverages Groq's infrastructure and Llama3 models to produce structured, markdown-formatted books with chapters, code examples, and tables in seconds.

How It Works

The application employs a scaffolded prompting strategy, utilizing a larger Llama3 model to generate the book's overall structure and chapter titles. Subsequently, a smaller Llama3 model is used to generate the content for each chapter, focusing solely on the chapter title for context. This dual-model approach aims to balance generation speed with content quality.

Quick Start & Requirements

• Hosted Version: Available at infinite.benjamin.sh
• Local Setup:
  1. Set GROQ_API_KEY environment variable (optional).
  2. Create and activate a Python virtual environment.
  3. Install dependencies: pip3 install -r requirements.txt
  4. (Windows) May require GTK3 installation: GTK for Windows
  5. Run the app: python3 -m streamlit run main.py
• Prerequisites: Python 3, Groq API Key.

Highlighted Details

• Generates entire books (e.g., 100 pages) in seconds.
• Outputs content in markdown, supporting tables and code blocks.
• Allows downloading the complete book content as a text file.
• Recent updates include inference statistics and PDF download functionality.

Maintenance & Community

• Active development with recent updates in May and June 2024.
• Contributions via Pull Requests are welcome.

Licensing & Compatibility

• The repository does not explicitly state a license in the provided README.

Limitations & Caveats

The generated content may be inaccurate or contain placeholders, and the tool is intended primarily for entertainment purposes. Future development aims to improve fiction generation by incorporating broader book context beyond just chapter titles.