An AI-powered tabletop RPG assistant that helps Game Masters create immersive Star Wars narratives using graph databases, vector search, and local language models.
- AI Story Generation: Context-aware narrative generation with streaming responses
- Multi-Database Integration: Neo4j (graph), MongoDB (documents), Weaviate (vectors)
- Local LLM: Privacy-focused with Mistral 7B running via LocalAI
- Web Interface: React 18 frontend with Star Wars-themed UI
- GPU Accelerated: Optimized for NVIDIA GPUs (RTX 4080 tested)
- Frontend: React 18 + TypeScript + Vite + Styled Components
- Backend: Node.js + Express + TypeScript
- Databases: Neo4j, MongoDB, Weaviate
- AI/ML: LocalAI with Mistral 7B
- Infrastructure: Docker Compose with GPU support
- Ubuntu 22.04+ or compatible Linux
- Docker & Docker Compose
- NVIDIA GPU with 8GB+ VRAM (optional but recommended)
- 32GB+ RAM
- 100GB+ free disk space
-
Clone the repository:
git clone https://github.com/magicat777/swrpg-gen.git cd swrpg-gen -
Copy environment template:
cp .env.example .env # Edit .env with your configuration -
Deploy services:
./scripts/deployment/deploy.sh
-
Access the application:
- Frontend: http://localhost:3001
- API: http://localhost:3000
Comprehensive documentation is available in the /docs directory:
- Development: 89% complete (Phase 8/13)
- Production Ready: Passed E2E testing
- Current Phase: Deployment & Source Code Integration
Contributions are welcome! Please read our contributing guidelines before submitting PRs.
This project is licensed under the MIT License - see the LICENSE file for details.
- LocalAI community for the excellent local LLM runtime
- Neo4j, MongoDB, and Weaviate teams for their databases
- The Star Wars fan community
May the Force be with your storytelling!