How khanhthanhdev/rag-application works
Main Components
a) Frontend Layer:
• Admin UI (port 8000/admin)
• Chat UI (port 8000)
• Management UI (for multiple RAGapps deployment)
b) Backend Services:
• RAGapp Core Service
• Manager Service (for orchestrating multiple RAGapps)
• Authentication Service (Keycloak integration)
c) External Integrations:
• AI Model Providers (OpenAI, Gemini, Ollama)
• Vector Databases (Qdrant, Chroma)
• API Gateway/Traefik (for routing and security)
Specific Components to Detail
Backend Structure:
Controllers -> Services -> Engine -> VectorDBs
• > Workflows
• > Models
Frontend Structure:
UI Components -> Client APIs -> Backend Routes
[Client Layer]
• ↓
[API Gateway/Security Layer]
• ↓
[Application Layer]
• ↓
[Services Layer]
• ↓
[Storage Layer]
Key Relationships to Highlight
• How the Manager service orchestrates multiple RAGapp instances
• Interaction between RAGapp and various AI providers
• Demonstrate the security flow through the API Gateway
• Vector database integration points
Important Technical Details
• Port mappings (8000)
• Docker container boundaries
• Configuration points (YAML files)
• API endpoints (/admin, /docs, /)
Deployment Variations
Two deployment scenarios:
Multiple RAGapps with management UI
This system design should emphasize the enterprise-ready nature of the application while showing its flexibility in supporting different AI models and deployment scenarios.