OpenManusJava

An Intelligent Thinking System Based on Java - A Multi-modal Agent Framework with Fast/Slow Thinking

🚀 Quick Start • 🎯 Features • 🏗️ Architecture •

📋 Project Overview

OpenManusJava is an intelligent thinking system developed based on Spring Boot and LangChain4j. It adopts a "fast thinking/slow thinking" dual-mode architecture, combining the efficiency of direct output with the depth of a think-execute-reflect loop. The system can automatically or manually select the most appropriate thinking mode according to task complexity, significantly improving the processing quality of complex tasks.

🎯 Features

🧠 Multi-modal Intelligent Thinking

• Fast Thinking Mode: Direct execution for high efficiency, suitable for simple tasks.
• Slow Thinking Mode: Think-execute-reflect loop, suitable for complex tasks.
• Automatic Mode: Intelligently selects the thinking mode based on task complexity.

💭 Intelligent Agent System

• FastThinkWorkflow: A workflow for quick responses.
• ThinkDoReflectWorkflow: A workflow for cyclic reflection.
• Multiple Specialized Agents: Specialized agents for thinking, execution, reflection, etc.

🔧 Tool Ecosystem

• Code Execution: Executes code and analyzes the results.
• File Operations: Manages files and content.
• Web Access: Intelligently retrieves information.

🎨 User Interface

• Modern 3-Column Workspace:
  • Left: An intelligent chat panel for core human-computer interaction.
  • Middle: A versatile tool panel displaying structured search results, tool outputs, and files.
  • Right: A browser workspace with multi-tab support, address bar navigation, and dual-mode (Web/VNC) capabilities.
• Real-time Thinking Process: Visualizes the AI's thinking steps and logs.
• Responsive Design: Adapts to desktop, tablet, and mobile devices.

🖼️ UI Preview

Note: Some websites block being embedded in an iframe via security headers like X-Frame-Options or CSP frame-ancestors. If you see “此网站无法在此预览”, enable the “代理” toggle in the address bar to load the page through the backend proxy.

🏗️ Architecture

Core Architecture Diagram

graph TD
    User --> UI[Web Interface]
    UI --> Controller[AgentController]
    
    Controller --> Service[AgentService]
    
    subgraph "Workflow"
        Service -->|Complex Task| TDR[ThinkDoReflectWorkflow<br/>Deep Thinking]
        Service -->|Simple Task| FT[FastThinkWorkflow<br/>Quick Response]
    end
    
    TDR --> TA[ThinkingAgent<br/>Analyze & Plan]
    TA --> EA[ExecutionAgent<br/>Execute Task]
    EA --> RA[ReflectionAgent<br/>Evaluate Result]
    RA -->|Task Complete| FinalResult[Final Result]
    RA -->|Needs More Work| TA
    
    subgraph "Tool Layer"
        CodeTool[Code Execution Tool]
        FileTool[File Operation Tool]
        SearchTool[Information Retrieval Tool]
    end
    
    EA --> CodeTool
    EA --> FileTool
    EA --> SearchTool
    
    FT --> FinalResult
    FinalResult --> WebSocket --> UI

Technology Stack

🚀 Quick Start

Prerequisites

• Java 21+
• Maven 3.9+
• Alibaba Cloud Bailian API Key (or other supported LLM service)

Installation

1. Clone the project

   git clone https://github.com/OpenManus/OpenManus-Java.git
   cd OpenManus-Java

2. Configure API keys (recommended) Copy dotenv.example to .env and fill in your API keys:

3. Optional: override application config Use src/main/resources/application-example.yml as a reference, and create src/main/resources/application-local.yml for personal overrides.

4. Start the application

   If you created application-local.yml, run with the local profile:

   mvn spring-boot:run -Dspring-boot.run.profiles=local

5. Access the service Open your browser and go to: http://localhost:8089

For a one-command setup (Docker Compose), see: docs/QUICK_START.md.

📊 Usage

Unified API Endpoint

All interactions are handled through a unified streaming API, think-do-reflect-stream, which automatically processes and returns real-time progress.

# Example Request
curl -X POST http://localhost:8089/api/agent/think-do-reflect-stream \
  -H "Content-Type: application/json" \
  -d '{"input": "Analyze the development trend of the tourism industry during the Spring Festival."}'

API Documentation

Swagger UI: http://localhost:8089/swagger-ui.html

📬 Contact Me

• WeChat: leochame007
• Email: liulch.cn@gmail.com

🙏 Acknowledgements

Thanks to the following open-source projects for their support:

• LangChain4j
• Spring Boot

📄 License

This project is licensed under the MIT License.

🌟 If this project is helpful to you, please give it a Star!