storytelling 0.2.13

AI-powered narrative generation system implementing Software 3.0 methodology - The Soul of Your Story's Blueprint

Storytelling: Turn Your Ideas Into Complete Narratives

Transform a simple story idea into a fully-written, multi-chapter narrative. WillWrite is an AI-powered storytelling partner that handles the heavy lifting—outline generation, chapter planning, scene development, and revision—so you can focus on your creative vision.

What It Does

Create complete stories in minutes. You provide a story prompt. The system generates:

• A structured outline with plot, characters, and themes
• Individual chapters with detailed chapter outlines
• Four polished scenes per chapter
• Multiple revision passes for consistency and quality
• Session persistence—pause and resume anytime

Powered by advanced AI orchestration. Built on LangGraph and LangChain with multi-LLM provider support, knowledge-base integration, and intelligent retrieval-augmented generation (RAG).

Getting Started

Installation

# Basic installation
pip install storytelling

# With full features (web fetching, local embeddings, cloud logging)
pip install storytelling[all]

Your First Story

# Create a prompt file
echo "A detective discovers a mysterious library" > my_prompt.txt

# Generate your story
storytelling --prompt my_prompt.txt --output my_story.md

Your completed story appears in my_story.md.

Using Your Own LLMs

# With local Ollama
storytelling --prompt prompt.txt \
  --initial-outline-model "ollama://mistral@localhost:11434"

# With Google Gemini
storytelling --prompt prompt.txt \
  --initial-outline-model "google://gemini-2.5-flash"

Key Features

Flexible AI Providers

• Google (Gemini, PaLM)
• Ollama (local models)
• OpenRouter (community models)
• Custom endpoints

Knowledge-Aware Generation

• Integrate your own knowledge bases
• Web content retrieval
• Semantic context enhancement
• Consistent worldbuilding

Session Management

• Pause and resume at any stage
• Checkpoint-based recovery
• Narrative branching
• Full generation history

Production Ready

• Comprehensive logging with Langfuse
• Type-safe with Pydantic
• Configurable for various use cases
• Extensible architecture

Command Line

storytelling --help                    # See all options
storytelling --prompt file.txt         # Generate story
storytelling --list-sessions           # See your past sessions
storytelling --resume <session-id>     # Continue a session

Documentation

• Full Setup Guide: See scripts/init.sh for environment setup
• Architecture: Check rispecs/ directory for detailed specifications
• Specifications: RISPECS.md contains implementation architecture
• Prompts & Models: rispecs/Prompts.md and rispecs/LLM_Provider_Specification.md

Development

# Setup development environment
./scripts/init.sh

# Run tests
make test

# Check code quality
make lint

# Build and release
make release-check

License

CC0-1.0 License - See LICENSE