AI GAL is a Ren'Py-based framework designed for generating unique Galgame stories and assets. It targets users seeking personalized visual novel experiences by leveraging AI for narrative, character art, voice, and music generation. The primary benefit is the creation of dynamic, user-driven narratives with AI-generated content, offering a novel approach to visual novel development and consumption.

How It Works

The project integrates multiple AI models within the Ren'Py engine. It utilizes GPT models for story generation (titles, outlines, dialogue, chapters), AI painting APIs (like Stable Diffusion) for visual assets (characters, backgrounds), and GPT-SOVITS for character voice synthesis. AI music generation based on emotional tags is also included. Users can influence the narrative through theme settings and branch selection, with options for both local and cloud-based AI processing to manage hardware demands.

Quick Start & Requirements

• Primary install: Launch via the provided "AI GAL Launcher".
• Prerequisites: Local installation of Stable Diffusion and GPT-SOVITS, and API access (or local deployment) for LLMs (e.g., Deepseek-R1). Cloud mode requires running Stable Diffusion and GPT-SOVITS APIs remotely.
• Links: Project Docs: https://tamikip.github.io/AI-GAL-doc/, Community QQ Group: 982330586.

Highlighted Details

• Supports ComfyUI for AI image generation (V1.6).
• Features multi-language support (English, Japanese) and a snapshot function for saving/sharing story progress (V1.5).
• Includes a dialogue mode (β) for one-on-one character interaction and user-customizable story modes (V1.6).
• Offers cloud modes for AI painting and voice generation, reducing local hardware demands (V1.1).

Maintenance & Community

The project receives regular updates, with the latest log (V1.6) dated March 19, 2025. A community QQ group (982330586) is available for support.

Licensing & Compatibility

The main project is licensed under MIT. However, it incorporates components under LGPL and Zlib licenses. Users must ensure all relevant licenses are included upon distribution. Compatibility for commercial use or linking with closed-source projects should be carefully reviewed due to the inclusion of LGPL components.

Limitations & Caveats

Dialogue mode is currently in β testing. User-selected story branches may result in slower content generation. The project relies on external AI services or significant local setup (Stable Diffusion, GPT-SOVITS, LLM APIs), which can be resource-intensive or require API key management.