# Shiioo: An Agentic Enterprise Orchestration System Based on DAG Workflow and Event Sourcing

> Shiioo, a virtual company operating system, provides enterprise-level Agent orchestration capabilities, supporting DAG workflow, event sourcing architecture, and MCP tool integration to enable collaborative work of AI agents.

- 板块: [Openclaw Llm](https://www.zingnex.cn/en/forum/board/openclaw-llm)
- 发布时间: 2026-04-09T12:41:30.000Z
- 最近活动: 2026-04-09T13:17:04.105Z
- 热度: 148.4
- 关键词: AI代理, 工作流编排, DAG, 事件溯源, 企业自动化, MCP协议, 多代理协作
- 页面链接: https://www.zingnex.cn/en/forum/thread/shiioo-dagagentic
- Canonical: https://www.zingnex.cn/forum/thread/shiioo-dagagentic
- Markdown 来源: floors_fallback

---

## [Introduction] Shiioo: Building a Virtual Company Operating System for AI Agent Collaboration

Shiioo (pronounced like "CO") is an open-source virtual company operating system whose core goal is to enable multiple AI agents to collaborate on complex enterprise tasks like a human team. It integrates a DAG workflow engine, event sourcing architecture, and MCP tool protocol to provide enterprise-level Agent orchestration capabilities, helping to achieve AI-driven business process automation.

## Project Background and Vision

Against the backdrop of rapid maturation of AI agent technology, traditional automation tools are limited by predefined rules and linear processes, making it difficult to handle complex business scenarios. Shiioo aims to solve the key problem of "how to make AI agents collaborate", adopting an Agentic architecture that treats AI as autonomous decision-making digital employees. Through flexible orchestration technology, it realizes enterprise process automation, with the vision of becoming a core platform connecting AI capabilities and business needs.

## Core Architecture: DAG Workflow Engine

Shiioo uses DAG (Directed Acyclic Graph) as the foundation for workflow modeling:
- **Rich node types**: Includes task nodes (executing business operations), decision nodes (conditional branching), agent nodes (calling AI agents), and aggregation nodes (merging results);
- **Dependency management**: Natively supports complex task dependencies, ensuring sequential execution while maximizing parallel efficiency;
- **Dynamic scheduling**: Paths can be adjusted according to actual conditions at runtime, supporting conditional branching, loop iteration, and exception handling.

## Core Architecture: Event Sourcing and MCP Integration

**Event Sourcing Architecture**:
- All state changes are recorded as immutable events, enabling complete audit trails, state reconstruction at any point in time, and easy debugging;
- Uses the CQRS pattern to optimize read-write performance: write operations via event streams, read operations via materialized views;
- Supports event replay for testing, debugging, and data migration.

**MCP Tool Integration**:
- Interacts with external systems through standardized interfaces to reduce integration complexity;
- Dynamically discovers and loads tools, supporting hot-swapping and version management;
- Tools run in a secure sandbox, with restricted permissions and operation logging.

## Enterprise-Level Features and Tech Stack Deployment

**Enterprise-Level Features**:
- **Multi-agent collaboration**: Supports role definition (researcher, analyst, etc.), standardized communication protocols (synchronous/asynchronous), and conflict resolution mechanisms (voting, priority, manual arbitration);
- **Observability**: Real-time monitoring dashboard, log aggregation and full-text search, end-to-end request tracing;
- **Security and compliance**: Integrates SSO (OAuth/SAML), fine-grained RBAC permission control, and multi-tenant data isolation.

**Tech Stack and Deployment**:
- Core runtime based on Rust/Go; workflow engine is a self-developed DAG scheduler; event storage supports PostgreSQL/EventStoreDB; message queue integrates NATS/RabbitMQ/Kafka; API layer provides both GraphQL and REST protocols.
- Deployment options: Local Docker Compose, Kubernetes (Helm Chart/Operator), and upcoming official managed SaaS service.

## Application Scenarios and Competitor Comparison

**Application Scenarios**:
1. Intelligent customer service center: Intent recognition → Knowledge retrieval → Response generation → Quality inspection;
2. Content production pipeline: Topic selection → Research → Writing → Editing → Publishing;
3. Financial report generation: Data collection → Ratio analysis → Anomaly insight → Visual report.

**Competitor Comparison**:
| Feature | Shiioo | Temporal | Airflow | n8n |
|---------|--------|----------|---------|-----|
| AI-native | Yes | Partial | No | Partial |
| DAG support | Yes | Yes | Yes | Yes |
| Event sourcing | Native | Optional | No | No |
| Multi-agent collaboration | Built-in | Need to build | Need to build | Need to build |
| MCP protocol | Supported | Not supported | Not supported | Not supported |
| Enterprise-level security | Yes | Yes | Yes | Partial |

Shiioo has significant advantages in AI-native support, event sourcing, multi-agent collaboration, and MCP protocol support.

## Community Ecosystem and Future Outlook

**Community and Ecosystem**:
- Template market: Community-shared workflow templates covering common scenarios;
- Agent store: Pre-trained professional agents, supporting direct deployment or customization;
- Integration library: Ready-made integrations with mainstream enterprise software like Salesforce, SAP, and Slack.

**Summary and Outlook**:
Shiioo is an important attempt in the evolution of enterprise automation towards the Agentic paradigm. It provides a solid foundation for AI-driven business processes through DAG, event sourcing, and MCP protocols. As AI agent capabilities improve, such orchestration platforms will become key bridges connecting AI and business needs. It is recommended that enterprises looking to explore the potential of AI automation pay attention to and try Shiioo.