This project provides a comprehensive AI-powered writing assistant focused on novel creation and general AI chat. It targets writers, content creators, and users interested in exploring AI for creative tasks, offering features like knowledge base integration, character and world-building tools, and multi-format content generation.

How It Works

The system leverages a modular architecture, integrating various AI models and local data sources. Key features include a "Knowledge Base" for custom data parsing, a "Creation Library" for structured writing workflows (plot, character, chapter generation), and a "Book Splitting Library" for analyzing and restructuring existing texts. It supports multiple AI platforms (OpenAI, Gemini, Zhipu, Ollama) and local GGUF models, allowing flexible backend configurations.

Quick Start & Requirements

• Install: pip install langchain==0.1.7
• Prerequisites: Python, local vectorization model (e.g., vector/m3e), potentially API keys for cloud AI services.
• Resources: Requires local disk space for models and data. Setup time varies based on model downloads and configuration.
• Docs: Feishu Docs Tutorial

Highlighted Details

• Supports AI group chat with custom role settings.
• Features a "Knowledge Graph" for managing entities and relationships.
• Offers semi-automatic writing workflows, from synopsis to polished text.
• Includes a "Book Splitting Library" with unlimited analysis and custom export options.

Maintenance & Community

The project is actively developed with frequent updates listed in the README. Community engagement appears to be primarily through WeChat Official Accounts ("多样分享") and Bilibili.

Licensing & Compatibility

The README does not explicitly state a license. The project mentions "public version" and "donation version" features, suggesting potential restrictions or tiered access. Commercial use compatibility is not specified.

Limitations & Caveats

The project is described as "experimental" and "for testing purposes." Some features, like specific domestic API integrations, are noted as having suboptimal performance. Users may need to obtain API keys or specific model files separately.