art-from-the-machine/Mantella

Mantella is a Skyrim and Fallout 4 mod which allows you to naturally speak to NPCs using a Speech-to-Text → LLMs → Text-to-Speech pipeline

What it solves

Mantella allows players to have natural, AI-driven conversations with NPCs in Skyrim and Fallout 4, replacing static dialogue trees with dynamic interactions.

How it works

The mod integrates several AI technologies to create a conversational loop: it uses speech-to-text (Moonshine or Whisper) to capture player input, Large Language Models (LLMs) to generate responses, and text-to-speech (Piper, xVASynth, or XTTS) to voice the NPC responses.

Who it’s for

Gamers and modders of Skyrim and Fallout 4 who want more immersive, open-ended interactions with in-game characters.

Highlights

  • Supports multiple speech-to-text engines (Moonshine, Whisper).
  • Compatible with various text-to-speech options (Piper, xVASynth, XTTS).
  • Works across both Skyrim and Fallout 4.