AI Dungeon Master for Tabletop RPGs

Project Overview

This project is an AI-assisted web application concept built around .NET application architecture, prompt engineering, state management, and cloud AI service integration. It serves as an AI Dungeon Master for tabletop RPG sessions, generating dynamic adventures, NPCs, dialogue, plot twists, and campaign context in real time.

Technical Scope

  • C# and ASP.NET Core application development
  • AI workflow design and prompt engineering
  • API integration with language models and speech services
  • SQL-backed state persistence and campaign data modeling
  • Asynchronous processing for user-facing AI features
  • Product thinking around user experience, reliability, and latency

Key Features

  1. Dynamic Adventure Generation: Uses AI to create unique, on-the-fly adventures tailored to player choices and party composition.
  2. NPC Creation and Interaction: Generates diverse NPCs with distinct personalities and backstories.
  3. Real-time Plot Twists: Introduces unexpected elements to keep the story engaging and challenging.
  4. Voice Synthesis for NPC Dialogue: Brings characters to life with AI-generated voices.
  5. Procedural Map Generation: Creates detailed, context-appropriate maps for various settings.
  6. Player Decision Impact Tracking: Analyzes and incorporates player choices to shape the ongoing narrative.

Technical Stack

  • Backend: C#, ASP.NET Core
  • AI Integration: OpenAI API-style language model workflows
  • Database: SQL Server (for campaign management and persistence)
  • Frontend: Blazor WebAssembly (for interactive UI components)
  • Cloud Services: Azure Cognitive Services for text-to-speech concepts

Implementation Details

AI Adventure Engine

  • Uses language model prompts to generate coherent and context-aware story elements.
  • Implements custom prompt engineering to keep generated content grounded in tabletop RPG rules, player context, and campaign history.
  • Maintains campaign state so generated content can respond to prior decisions instead of acting like isolated one-off prompts.

Voice Synthesis

  • Integrates with Azure Cognitive Services for Text-to-Speech capabilities.
  • Matches voice characteristics to NPC attributes for immersive dialogue delivery.

Procedural Map Generation

  • Custom algorithm for creating varied and logical dungeon layouts, town maps, and wilderness terrains.
  • Integrates with the story generation to ensure map relevance to the current adventure.

Player Choice Analysis

  • Tracks player decisions and party actions.
  • Uses structured state and AI-generated reasoning to guide future story direction and generate appropriate content.

Development Challenges and Solutions

  1. Maintaining Narrative Coherence: Implemented a context management system to ensure AI-generated content remains consistent throughout the campaign.
  2. Performance Optimization: Used caching strategies and asynchronous processing to handle real-time AI responses without impacting user experience.
  3. Balancing AI Creativity with D&D Rules: Developed a rules-checking module to validate AI-generated content against D&D 5e ruleset.

Future Enhancements

  • Integration with virtual tabletop platforms for seamless online play.
  • Expansion of the AI model to include more game systems beyond D&D.
  • Implementation of a collaborative storytelling mode where multiple human players can contribute to the narrative alongside the AI.

This project showcases backend development, AI service integration, prompt engineering, stateful application design, SQL-backed persistence, and the ability to translate an open-ended user experience into a structured technical system.