Skip to main content

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:

ItemMinimum Requirement
CPU>= 2 Core
Memory>= 4 GiB
OSmacOS / Linux / WSL2
DockerDocker Engine 20.10+
Docker Composev2.0+

Architecture

TGO uses a microservices architecture with the following core services:

ServiceDescription
tgo-apiBackend API service for business logic
tgo-aiAI inference service for agent conversations
tgo-ragRAG service for knowledge base retrieval
tgo-platformPlatform management service
tgo-webWeb management console
tgo-widgetEmbeddable customer service widget

Next Steps