1. Project Overview

Wegent

An Open-source Platform for Defining, Organizing, and Running Intelligent AI Entities

Build Your AI Intelligent Entity Work Team

From Programming Assistants to News Analysis - Deploy Truly Capable Intelligent Agents

Wegent is an Open-source AI Native Operating System that Enables Massively Defining, Organizing, and Running Intelligent Agents. Based on Kubernetes-style Declarative API and CRD (Custom Resource Definition) Design Patterns, Wegent Provides a Standardized Framework for Creating and Managing AI Intelligent Entity Ecosystems

2. Open Source License

Licensed under Apache-2.0

3. Screenshots

4. Feature Overview

Core Capabilities

1. Configuration-Driven Intelligent Entity Teams
   Define and Run Customized Agent Teams through YAML Configuration, Providing Web UI without Secondary Development
2. Multi-Engine Architecture
   Supporting Agno and Claude Code Two Agent Execution Engines at the Bottom and Dialog and Coding Two Modes at the Top
3. Independent Sandbox Environment
   Each Agent Team Runs in an Independent Sandbox Environment, Supporting Multiple Agent Teams Running Simultaneously
4. Advanced Collaboration Patterns
   Dialog Mode Can Implement Parallel, Leader, etc. Agent Collaboration Patterns, Completing Complex Workflows such as News Insight and Content Retrieval
5. AI Coding Integration
   Coding Mode Can Integrate with GitHub/GitLab, etc. Code Services, Implementing AI Coding Workflows such as Code Development and Review

Core Concepts

• Ghost
  The 'Soul' of Intelligent Entities - Defining Personalities, Abilities, and Behavior Patterns
• Model
  AI Model Configuration - Defining Environment Variables and Model Parameters
• Shell
  Executable Programs - Able to Launch Intelligent Entities
• Bot
  Complete Intelligent Entity Instances, Combining Ghost + Shell + Model
• Team
  Composed of Multiple Bots + Collaboration Models, Defining How Intelligent Entities Cooperate
• Collaboration
  Defining Interaction Patterns between Bots in the Team (Similar to Workflows)
• Workspace
  Isolated Work Environments for Tasks and Projects
• Task
  Executable Work Units Assigned to Teams

Detailed YAML Configuration Documents:

• Complete YAML Configuration Examples and Field Descriptions

What Can You Build?

Wegent Lets You Create Powerful AI Applications through Intelligent Entity Orchestration:

Web-based Programming Assistants

Building a Full-featured Development Environment in the Browser, Integrating with GitHub, Supporting Local or Cloud-based Independent Development Spaces, and Running Multiple Coding Agents Simultaneously

News Intelligent Platforms

Creating Intelligent News Aggregation and Analysis Systems

Customizable Intelligent Entity Applications

Infinite Possibilities - Building Intelligent Entities for the Following Scenarios:

• Data Analysis
  Automated Report Generation and Visualization
• Content Creation
  Blog Articles, Social Media, and Marketing Materials
• Customer Support
  Intelligent Chatbots with Contextual Understanding
• DevOps Automation
  CI/CD Pipeline Management and Monitoring
• Research Assistants
  Literature Reviews and Knowledge Synthesis

5. Tech Stack

Architecture

Quick Start

Pre-requisites

• Docker and Docker Compose
• Git

1. Clone the Repository

git clone wegent
cd wegent

2. Start the platform

docker-compose up -d

3. Access the Web interface

• Open http://localhost:3000 in your browser

4. Configure GitHub access token

• Configure your GitHub access token according to the page instructions

5. Configure Bot

• Wegent has a built-in development Bot. For Claude Code runtime, set the following environment variables:

{
  "env": {
    "ANTHROPIC_MODEL": "openrouter,anthropic/claude-sonnet-4",
    "ANTHROPIC_AUTH_TOKEN": "sk-xxxxxx",
    "ANTHROPIC_BASE_URL": "http://xxxxx",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "openrouter,anthropic/claude-haiku-4.5"
  }
}

Important note - environment variable naming:

Different Shell runtimes use different environment variable names for API authentication:

• Claude Code Shell
  Use ANTHROPIC_AUTH_TOKEN
• Agno Shell
  Use ANTHROPIC_API_KEY

Please set the correct environment variable according to your Shell configuration. For specific requirements, please refer to the Shell documentation or executor/agents/ code.

6. Run tasks

• On the task page, select your project and branch, describe your development requirements, such as implementing the bubble sort algorithm using Python