RachelForster/Shinsekai

What it solves

Shinsekai is a desktop assistant designed for Galgame, Otome, and story-driven RPGs. It solves the problem of fragmented workflows by integrating LLM-driven dialogue, character art (sprites) with emotional synchronization, text-to-speech (TTS), and speech-to-text (ASR) into a single, cohesive experience.

How it works

The system splits its functionality into two primary interfaces: a React-based settings center for managing APIs, characters, and plugins, and a dedicated chat window for the actual performance. It connects to various LLM providers via API, supports multiple TTS engines (like GPT-SoVITS and Genie TTS), and can integrate image generation via ComfyUI. It uses a plugin SDK and the Model Context Protocol (MCP) to extend its capabilities with external tools and services.

Who it’s for

It is aimed at creators and players of visual novels and story-driven games who want to create immersive AI-driven character interactions with synchronized visuals and audio.

Highlights

• Integrated Performance Pipeline: Combines chat templates, session history, emotional sprite switching, and voice synthesis in one workflow.
• Extensible Architecture: Features a plugin SDK and MCP support to add new LLM, TTS, ASR, and T2I adapters or external tools.
• Emotional Linkage: Automatically switches character sprites based on emotion tags in the AI's dialogue.
• Long-term Memory: Manages and retrieves character-specific memories across different sessions to maintain plot consistency.
• Customizable Visuals: Includes a built-in theme manager for editing chat skins, fonts, and stage styles.
• Character Library: Provides over 100 pre-made character packs for immediate use.