Mantella enhances Skyrim and Fallout 4 by enabling natural, AI-powered conversations with NPCs. It targets players seeking deeper immersion, leveraging speech-to-text, large language models, and text-to-speech technologies to create dynamic dialogue interactions.

How It Works

Mantella integrates Whisper for speech-to-text, a chosen LLM for generating NPC responses, and xVASynth/XTTS for text-to-speech. This pipeline allows players to speak naturally into their microphone, have their speech transcribed, processed by an LLM to generate a contextually relevant response, and then spoken aloud by an AI voice, creating a seamless conversational experience.

Quick Start & Requirements

• Install via pip install -r requirements.txt after cloning the repository.
• Requires Python 3.11 and a GPT secret key stored in GPT_SECRET_KEY.txt.
• Configuration is managed via config.ini.
• Official documentation is available on the gh-pages branch.
• Discord support is available for setup assistance.

Highlighted Details

• Enables natural voice interaction with game NPCs.
• Supports multiple text-to-speech engines (Piper, xVASynth, XTTS).
• Leverages Whisper for speech-to-text.
• Requires integration with a Large Language Model.

Maintenance & Community

• The project is actively developed, with related repositories for specific game versions (Skyrim, Fallout 4, Fallout 4 VR).
• Community support is available via Discord.

Licensing & Compatibility

• Uses material licensed under the Creative Commons Attribution-Share Alike License.
• Source code is included, suggesting potential for community contributions.

Limitations & Caveats

This development version is explicitly stated as prone to error and not recommended for general use; users are advised to seek the latest stable release. Updates across related repositories are often intertwined, requiring careful version management.