Skip to main content

Introduction to TGO

TGO is an open-source AI agent customer service platform. We are dedicated to helping enterprises build smarter and more efficient customer service systems through the Multi-Agents collaboration model.

Unlike traditional chatbots, TGO emphasizes collaboration between agents. You can assemble a team composed of agents with different capabilities to collaboratively handle complex customer inquiries, after-sales support, and business lead generation.

Key Highlights

  • 🤖 Multi-Agent Orchestration: Flexibly create and orchestrate multiple AI agents, allowing them to perform their respective duties and work together.
  • 📚 Knowledge Base Enhancement (RAG): Built-in high-performance RAG system supporting PDF, Docx uploads, web crawling, etc., making AI understand your business.
  • 🛠️ MCP Tool Extension: Supports Model Context Protocol, easily connecting to external APIs, databases, and business tools.
  • 🌐 Omnichannel Access: One system supporting Web, API, WeChat, Lark, DingTalk, and other mainstream channels.
  • 🤝 Human-AI Collaboration: Agents handle repetitive inquiries and seamlessly transfer to human agents at critical moments to improve customer satisfaction.

Technical Architecture

TGO adopts a modern microservices architecture to ensure system stability and scalability.

TGO Architecture

ServiceDescription
tgo-apiCore business service, handling logic for accounts, conversations, permissions, etc.
tgo-aiAgent orchestration center, responsible for LLM calls and agent scheduling.
tgo-ragKnowledge base retrieval service, handling document vectorization and semantic search.
tgo-platformExternal platform integration service, responsible for connecting to DingTalk, Lark, WeChat Work, etc.
tgo-webCustomer service management console, a modern dashboard built with React.
tgo-widgetLightweight visitor-side component, can be integrated into any webpage in minutes.
WuKongIMHigh-performance real-time messaging system, ensuring real-time message delivery.

System Requirements

Before deploying TGO, please ensure your environment meets the following requirements:

ItemMinimumRecommended
CPU2 Core4 Core+
Memory8 GiB16 GiB+
OSLinux / macOS / WSL2Ubuntu 22.04 LTS
EnvironmentDocker 20.10+Docker Compose v2.0+

Next Steps