Summary Shinsekai streamlines narrative RPG and Galgame creation by unifying LLM-driven character dialogue, dynamic character art (立绘) linkage, and integrated multimedia extensions (TTS, ASR, T2I) within a local desktop assistant. Targeting creators of story-centric games, it offers a cohesive workflow, reducing tool fragmentation and enhancing presentation focus.

How It Works The application features a dual-window design: a settings center for API, character, and plugin management, and a chat window for dialogue演出. Core components include LLM-driven interactions, emotion-tagged character art, and integrated TTS/ASR. Supports flexible, multi-model backends for LLMs, TTS, and T2I, allowing selection of preferred engines.

Quick Start & Requirements Installation options include source (git clone, pip install -r requirements.txt) or pre-built bundles from Releases. Developers recommend Python 3.10+ in a virtual environment. Bundles use platform-specific scripts (install.bat/.command, start.bat/.command) for Windows/macOS. Initial setup (API config, character import) takes ~5 minutes. Key documentation links are provided.

Highlighted Details

• Unified Workflow: Integrates character management, dialogue templating, session history, dynamic character art with emotion linkage, TTS/ASR, and T2I generation.
• Flexible Backends: Supports diverse LLM providers (OpenAI compatible), TTS engines (GPT-SoVITS, Genie TTS), and T2I workflows (ComfyUI) via configurable API settings.
• Extensible Architecture: A plugin system extends LLM tools, TTS/ASR/T2I adapters, and UI components. MCP protocol integrates external services as LLM tools.
• Local Data Management: Configuration, character data, and history reside locally in data/, simplifying backup and customization.

Maintenance & Community The project welcomes Issues and PRs. A community plugin registry (Shinsekai-Plugin-Registry) is available. Links to user and developer guides are provided.

Licensing & Compatibility The project's specific open-source license is not explicitly stated. Redistribution of associated assets requires adherence to their respective authors' licenses. Commercial use compatibility is undetermined due to the unclear project license.

Limitations & Caveats The primary README is in Chinese; an English version is linked. Advanced features like T2I may benefit from specific hardware. The unclear project license requires clarification for adoption decisions.