🌾 Stardew Sage - AI-Powered Stardew Valley Assistant
An intelligent AI agent system that brings the world of Stardew Valley to life through conversational AI, RAG vector databases, and multimodal analysis.
Features • Architecture • Quick Start • Documentation • Contributing
📖 Overview
Stardew Sage is an innovative AI-powered assistant designed for Stardew Valley fans. It combines the power of modern AI technologies to create an interactive experience where users can:
- 🎂 Get daily birthday stories - Discover which Stardew Valley character is celebrating their birthday with engaging AI-generated narratives
- 🎯 Receive daily missions - Get personalized daily tasks and challenges inspired by the game
- 🦐 Query game database - Ask questions about game mechanics, items, characters, and get intelligent responses powered by MCP (Model Context Protocol)
- 📚 RAG-powered search - Leverage vector database search for semantic understanding of game content
- 🖼️ Image analysis - Upload game screenshots and get AI-powered analysis and insights using multimodal vision models
✨ Features
🧠 AI Agent System
- Intelligent Conversational Agent: Built with Spring AI and Model Context Protocol (MCP) for context-aware responses
- Multi-Prompt Strategy: Different specialized prompts for birthdays, missions, database queries, and RAG searches
- Streaming Responses: Real-time SSE (Server-Sent Events) streaming for smooth user experience
🗄️ Data Management
- Comprehensive Game Database: MySQL database with 20+ tables covering all aspects of Stardew Valley:
- Characters (NPCs) and their preferences
- Items (crops, cooking, crafting, minerals, fish, etc.)
- Game mechanics (seasons, events, quests)
- Animals, buildings, and locations
- RAG Vector Store: Redis-based vector database for semantic search using embeddings
- Flexible Data Upload: API endpoints for uploading and managing vector embeddings
🎨 Modern Web Interface
- Vue.js Frontend: Responsive, game-themed UI with Stardew Valley aesthetics
- Real-time Updates: SSE integration for live AI response streaming
- Multiple Modules:
- MCP Chat Interface
- RAG-powered Q&A
- Image Analysis Tool
- Data Management Dashboard
🔧 Technical Capabilities
- Spring AI Integration: Leverages Spring AI framework for AI operations
- OpenAI-Compatible APIs: Works with various LLM providers (GLM-4.5, GPT, etc.)
- Vector Embeddings: Text embedding generation and similarity search
- Multimodal Support: Vision model integration for image understanding
- MyBatis Plus: Efficient database operations with automatic code generation
🏗️ Architecture
Stardew-Sage/
├── aurora-admin/ # Backend admin module (RuoYi framework)
│ └── Spring Boot application for system management
├── aurora-mcp/ # AI MCP Server module
│ ├── Controller # REST API endpoints
│ ├── Service # Business logic and AI integration
│ ├── Model # Database entities (Stardew Valley data)
│ ├── Mapper # MyBatis database mappers
│ └── Prompt # AI prompt templates
├── aurora-ui/ # Frontend Vue.js application
│ ├── src/
│ │ ├── views/ # Vue components
│ │ │ ├── stardew/ # Stardew-specific pages
│ │ │ │ ├── mcp/ # MCP chat interface
│ │ │ │ ├── rag/ # RAG Q&A interface
│ │ │ │ ├── image/ # Image analysis
│ │ │ │ └── show/ # Data display
│ │ ├── api/ # API service layer
│ │ └── assets/ # Static resources
└── database.sql # Complete database schema and data
Technology Stack
Backend:
- Spring Boot 3.4.3
- Spring AI 1.0.0-M8 (MCP Server, OpenAI integration)
- MyBatis Plus 3.5.5
- MySQL 8.0
- RedisStack (Vector Store with RediSearch)
- Maven
Frontend:
- Vue.js 2.6.12
- Element UI 2.15.14
- Axios
- Server-Sent Events (SSE)
AI/ML:
- Model Context Protocol (MCP)
- OpenAI-compatible API support
- Vector embeddings (2048 dimensions)
- GLM-4.5V vision model
- RedisStack vector store for RAG (RediSearch module)
🚀 Quick Start
Prerequisites
- Java 17+ (for aurora-mcp)
- Java 8+ (for aurora-admin)
- Node.js 8.9+ and npm
- MySQL 8.0+
- RedisStack (Redis with RediSearch module for vector store)
- AI API Key (OpenAI, GLM, or compatible provider)
1. Database Setup
# Create and import database
mysql -u root -p
source database.sql
# This creates the 'ry-vue' database with all Stardew Valley data
Install RedisStack:
RedisStack is required for vector search capabilities (RediSearch module).
# Option 1: Using Docker (Recommended)
docker run -d --name redis-stack -p 6379:6379 -p 8001:8001 redis/redis-stack:latest
# Option 2: Install directly on Linux
curl -fsSL https://packages.redis.io/gpg | sudo gpg --dearmor -o /usr/share/keyrings/redis-archive-keyring.gpg
echo "deb [signed-by=/usr/share/keyrings/redis-archive-keyring.gpg] https://packages.redis.io/deb $(lsb_release -cs) main" | sudo tee /etc/apt/sources.list.d/redis.list
sudo apt-get update
sudo apt-get install redis-stack-server
# Option 3: On macOS with Homebrew
brew tap redis-stack/redis-stack
brew install redis-stack
# Verify RedisStack is running
redis-cli PING
# Should return: PONG
2. Backend Setup - MCP Server (aurora-mcp)
# Navigate to MCP server directory
cd aurora-mcp
# Configure application settings
# Edit src/main/resources/application.yml:
# - Database connection (MySQL)
# - RedisStack connection
# - AI API key and base URL
# Example configuration:
# spring:
# datasource:
# url: jdbc:mysql://localhost:3306/ry-vue
# username: root
# password: your_password
# data:
# redis:
# host: localhost
# port: 6379 # RedisStack default port
# ai:
# openai:
# api-key: your_api_key
# base-url: your_api_endpoint
# vectorstore:
# redis:
# index-name: stardew_index
# initialize-schema: true
# Build and run
mvn clean install
mvn spring-boot:run
# Server starts on port 8099
3. Backend Setup - Admin Server (aurora-admin)
# Navigate to admin directory
cd aurora-admin
# Configure application settings
# Edit src/main/resources/application.yml
# Build and run
cd ..
mvn clean install
cd aurora-admin
mvn spring-boot:run
# Server starts on port 8080
4. Frontend Setup (aurora-ui)
# Navigate to UI directory
cd aurora-ui
# Install dependencies
npm install
# For faster installation in China:
# npm install --registry=https://registry.npmmirror.com
# Start development server
npm run dev
# Application runs on http://localhost:80
5. Access the Application
- Frontend: http://localhost:80
- MCP API: http://localhost:8099
- Admin API: http://localhost:8080
Default admin credentials (if using RuoYi system):
- Username:
admin - Password:
admin123
📚 Documentation
API Endpoints
MCP Agent APIs (Port 8099)
Birthday Story Stream
GET /api/agent/getBirthInfoStream
Content-Type: text/event-stream
Returns: SSE stream with birthday character story
Daily Mission Stream
GET /api/agent/getMissionInfoStream
Content-Type: text/event-stream
Returns: SSE stream with daily mission
Database Query Stream
GET /api/agent/getInfoDetail?question=<your_question>
Content-Type: text/event-stream
Returns: SSE stream with query results using MCP tools
RAG Query Stream
GET /api/agent/getInfoRagDetail?question=<your_question>
Content-Type: text/event-stream
Returns: SSE stream with RAG-enhanced query results
Image Analysis
POST /api/image/analyze
Content-Type: multipart/form-data
Parameters:
- file: Image file (JPEG/PNG, max 5MB)
- question: Optional analysis question
Returns: Stream with image analysis
Resource Management APIs
Get All Tables
GET /api/table/list
Returns: List of all database tables
Get Table Info
GET /api/table/info?tableName=<table_name>
Returns: Table structure and data
Upload to RAG
POST /api/table/upload
Body: { tableName, columnSimpleMetas }
Returns: Upload status
Database Schema
The project includes 20+ tables for Stardew Valley data:
stardew_people- NPCs and charactersstardew_people_favorite- Character gift preferencesstardew_crop- Crop informationstardew_food,stardew_food_material- Cooking recipesstardew_fish,stardew_fish_task- Fishing datastardew_mineral,stardew_antique- Collectiblesstardew_animal,stardew_husbandry- Farm animalsstardew_craft,stardew_craft_material- Crafting- And many more...
Vector Store
The system uses RedisStack as a vector database for RAG:
- Index Name:
stardew_index - Embedding Model: embedding-3
- Dimensions: 2048
- Purpose: Semantic search across game content
- Requirements: RedisStack with RediSearch module for vector similarity search
🎮 Usage Examples
Example 1: Ask About a Character
// Frontend code example
import { getInfoDetailSSE } from '@/api/ai'
const eventSource = getInfoDetailSSE('Tell me about Abigail')
eventSource.onmessage = (event) => {
console.log(event.data) // Streaming response
}
Example 2: Upload Data to RAG
import { uploadToRAG } from '@/api/ai'
const result = await uploadToRAG('stardew_crop', [
{ columnName: 'name_ch', isChose: true },
{ columnName: 'remark', isChose: true }
])
Example 3: Analyze Game Screenshot
import { analyzeImage } from '@/api/ai'
const file = document.querySelector('input[type="file"]').files[0]
const response = await analyzeImage(file, 'What crop is this?')
🔧 Configuration
AI Model Configuration
Edit aurora-mcp/src/main/resources/application.yml:
spring:
data:
redis:
host: localhost # RedisStack host
port: 6379 # RedisStack port
ai:
openai:
base-url: https://open.bigmodel.cn/api/paas/v4/ # Your AI provider
api-key: your_api_key_here
chat:
options:
model: GLM-4.5V # Change to your preferred model
embedding:
options:
model: embedding-3
dimensions: 2048
vectorstore:
redis:
index-name: stardew_index
initialize-schema: true
Note: RedisStack is required for vector store functionality. Standard Redis does not include the RediSearch module needed for vector similarity search.
Frontend API Configuration
Edit aurora-ui/vue.config.js:
const baseUrl = 'http://localhost:8080' // Backend URL
const port = 80 // Frontend port
🤝 Contributing
Contributions are welcome! Here's how you can help:
- Fork the repository
- Create a feature branch (
git checkout -b feature/AmazingFeature) - Commit your changes (
git commit -m 'Add some AmazingFeature') - Push to the branch (
git push origin feature/AmazingFeature) - Open a Pull Request
Development Guidelines
- Follow existing code style and conventions
- Add comments for complex logic
- Update documentation for new features
- Test thoroughly before submitting PR
📝 License
This project is licensed under the MIT License - see the LICENSE file for details.
🙏 Acknowledgments
- Stardew Valley by ConcernedApe for the amazing game
- RuoYi framework for the admin system foundation
- Spring AI team for the excellent AI integration framework
- OpenAI and GLM for AI model capabilities
- The Stardew Valley community for game data and inspiration
Contact & Support
- GitHub Issues: For bug reports and feature requests
- Discussions: For questions and community support
Made with ❤️ for the Stardew Valley community
⭐ Star this repository if you find it helpful!