Introduction to TGO
TGO is an open-source AI agent customer service platform, dedicated to helping enterprises "build agent teams to serve customers".
Core Features
- Multi-Agent Orchestration: Create and manage multiple AI agents to collaboratively handle customer inquiries
- Knowledge Base (RAG): Built-in RAG system supporting documents, web pages, and various knowledge sources
- MCP Tool Integration: Support for Model Context Protocol to extend agent capabilities
- Multi-Channel Access: Support for web embedding, API, and various access methods
- Human Agent Collaboration: Seamless switching between AI agents and human agents
System Requirements
Deploying TGO requires the following minimum configuration:
| Item | Minimum Requirement |
|---|---|
| CPU | >= 2 Core |
| Memory | >= 4 GiB |
| OS | macOS / Linux / WSL2 |
| Docker | Docker Engine 20.10+ |
| Docker Compose | v2.0+ |
Architecture
TGO uses a microservices architecture with the following core services:
| Service | Description |
|---|---|
| tgo-api | Backend API service for business logic |
| tgo-ai | AI inference service for agent conversations |
| tgo-rag | RAG service for knowledge base retrieval |
| tgo-platform | Platform management service |
| tgo-web | Web management console |
| tgo-widget | Embeddable customer service widget |
Next Steps
- Quick Deploy - Deploy TGO on your server quickly
- Environment Variables - Learn how to configure the system
- Source Deployment - Build and deploy from source