Skip to main content
Back to Projects

CritChat

A mobile-first TTRPG social platform with AI-powered character interactions and intelligent group matching for tabletop role-playing game communities.

Flutter Dart Firebase Weaviate OpenAI API BLoC Pattern Firestore Firebase Auth Realtime Database Cloud Functions

Overview

CritChat is a revolutionary mobile-first social platform designed specifically for Tabletop Role-Playing Game (TTRPG) communities. Combining Snapchat-style ephemeral content with AI-powered character interactions, it provides a dedicated space for tight-knit campaign parties to connect, organize, and bring their characters to life through intelligent conversation systems.

Target User & Use Cases

Primary Users: TTRPG players, Game Masters, campaign groups, and players seeking new gaming communities.

Key Use Cases:

  • Character Interaction: AI-powered character conversations using RAG (Retrieval-Augmented Generation) system
  • Group Formation: Smart “Looking for Group” (LFG) system with hybrid matching algorithms
  • Campaign Organization: Fellowship management with real-time chat and coordination tools
  • Content Sharing: Ephemeral stories and persistent session recaps
  • Community Building: Friend systems and social discovery for TTRPG enthusiasts

Core Features Implemented

Revolutionary AI-Powered Character System

  • RAG Integration: Complete vector database system using Weaviate for character memory storage
  • @as Command System: Chat with characters using @as <character> <message> commands
  • Character Memory: Automatic vectorization and contextual retrieval of character experiences
  • Personality-Driven Responses: OpenAI GPT-4o-mini generates responses matching character personalities
  • Memory Management: Seamless addition and categorization of character backstory and experiences

Intelligent Group Discovery

  • LFG (Looking For Group) System: “Call to Adventure” posts with comprehensive matching
  • Hybrid Matching Algorithm: Combines semantic similarity (30%) with algorithmic compatibility (70%)
  • Smart Filtering: Priority weighting for play style (40%), game system (25%), campaign length (20%), schedule (10%), session format (5%)
  • Interest Management: “Answer Call” functionality with state tracking and visual feedback
  • Fellowship Formation: Seamless transition from LFG connections to active gaming groups

Real-Time Social Features

  • Fellowship Management: Campaign group creation and member coordination
  • Live Chat System: Real-time messaging with character integration support
  • Friend System: Social connections with discovery and invitation features
  • Notification System: Comprehensive real-time updates for all platform interactions
  • Polls & Voting: Democratic decision-making tools for session planning

Gamification & Engagement

  • XP System: Experience points for platform engagement and quality contributions
  • Level Progression: Character and user advancement tracking
  • Content Rewards: Bonus XP for session recaps and quality interactions
  • Achievement System: Recognition for platform participation and community building

Technical Implementation

Developed using AI-first engineering principles with a production-ready mobile architecture:

  • Frontend: Flutter with Material 3 design for cross-platform mobile experience
  • State Management: BLoC pattern for reactive, testable state management
  • Navigation: Go Router for declarative navigation and deep linking
  • Authentication: Firebase Auth with comprehensive user profile management
  • Database: Firestore for structured data, Firebase Realtime Database for chat
  • AI Integration: OpenAI API for LLM responses and text embeddings
  • Vector Database: Weaviate for character memory storage and similarity search
  • Backend Logic: Firebase Cloud Functions for serverless business logic
  • Testing: Comprehensive unit and widget tests with mocktail for mocking

Advanced AI Architecture

RAG (Retrieval-Augmented Generation) System

  • Vector Storage: Weaviate database with dedicated collections for character memories and LFG posts
  • Embedding Generation: OpenAI text-embedding-3-small for high-quality vector representations
  • Context Assembly: Character profile + retrieved memories + recent chat history for LLM prompts
  • Response Generation: Personality-consistent character responses with speech pattern matching
  • Graceful Degradation: Mock services for development with seamless production AI integration

Character Memory Pipeline

User adds memory → Automatic vectorization → Weaviate storage → Similarity search → Context retrieval → AI response generation

LFG Matching Architecture

User creates post → Vector embedding → Hybrid scoring → Compatibility ranking → Smart filtering → Match presentation

Results & Impact

  • AI-First Development: Complete production-ready platform built using AI-assisted engineering in 6-week development cycle
  • Revolutionary Architecture: First-of-its-kind RAG system integration in mobile TTRPG social platform
  • Production Infrastructure: Scalable Firebase backend with vector database integration
  • Character Interaction Innovation: AI-powered character conversations with persistent memory systems
  • Community Platform: Complete social ecosystem for TTRPG players from character creation to group coordination

This project represents a breakthrough in AI-integrated social gaming platforms, demonstrating how advanced machine learning techniques can enhance creative collaboration and character-driven storytelling in gaming communities. The combination of Flutter’s mobile-first architecture with cutting-edge AI technologies creates an unprecedented platform for TTRPG social interaction.

More Projects

View All Projects Get in Touch