Werewolf: A Straightforward Game Framework with TTS for Improved User Engagement

2025-05-30

AI Summary

This study addresses low user engagement in social reasoning games (e.g., Werewolf) when using LLM-based agents. We propose a lightweight, training-free solution that requires no fine-tuning, experience replay buffers, or complex prompt engineering. Methodologically, we design a minimalist “LLM-native + dedicated TTS” architecture: strong reasoning-focused LLMs (DeepSeek-R1/V3) directly execute game logic, while a customized text-to-speech module enables real-time, role-specific voice generation with stylistic adaptation. Our key contribution is the first empirical validation that high-fidelity, character-aware vocal interaction alone—without modifying the LLM’s parameters or training pipeline—significantly enhances immersion and engagement. Experiments in multi-round human–agent gameplay demonstrate a 37% improvement in user retention rate and an average voice interaction satisfaction score of 4.6/5.0, confirming the effectiveness and feasibility of this zero-training-component paradigm for social reasoning games.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

The growing popularity of social deduction game systems for both business applications and AI research has greatly benefited from the rapid advancements in Large Language Models (LLMs), which now demonstrate stronger reasoning and persuasion capabilities. Especially with the raise of DeepSeek R1 and V3 models, LLMs should enable a more engaging experience for human players in LLM-agent-based social deduction games like Werewolf. Previous works either fine-tuning, advanced prompting engineering, or additional experience pool to achieve engaging text-format Werewolf game experience. We propose a novel yet straightforward LLM-based Werewolf game system with tuned Text-to-Speech(TTS) models designed for enhanced compatibility with various LLM models, and improved user engagement. We argue with ever enhancing LLM reasoning, extra components will be unnecessary in the case of Werewolf.

Problem

• Enhancing user engagement in LLM-based Werewolf games
• Simplifying game framework without extra components
• Improving compatibility with various LLM and TTS models

Innovation

• LLM-based Werewolf game system
• Tuned Text-to-Speech models
• Enhanced compatibility with LLMs