TradingAgents-MCPmode

# TradingAgents-MCPmode Multi-Agent Trading Analysis System Based on MCP Tool ## 🌟 Project Overview **TradingAgents-MCPmode** is an innovative multi-agent trading analysis system that integrates the Model Context Protocol (MCP) tools to achieve intelligent stock analysis and trading decision-making processes. The system provides comprehensive market analysis, investment recommendations, and risk management through the collaboration of 15 specialized agents. ### 🎯 Core Features - **🤖 Multi-Agent Collaboration**: 15 specialized agents working together - **⚡ Parallel Processing**: The analyst team employs a parallel architecture, significantly enhancing analysis efficiency - **🔧 MCP Tool Integration**: Supports external data sources and real-time information retrieval - **📊 Comprehensive Analysis**: Seven-dimensional analysis covering company overview, market, sentiment, news, fundamentals, shareholder structure, and product business - **💭 Intelligent Debate**: Bullish/Bearish researcher debate mechanism with configurable debate rounds - **⚠️ Risk Management**: Three-tier risk analysis and management decision-making, supporting dynamic risk debates - **🎛️ Agent Control**: Frontend allows dynamic enabling/disabling of specific agents, flexibly customizing workflows - **🌀 Debate Round Configuration**: Real-time setting of investment and risk debate rounds on the frontend, precisely controlling analysis depth - **🔧 Flexible Configuration**: Control agent MCP permissions through environment variables - **🌍 Multi-Market Support**: US stocks (US), A-shares (CN), Hong Kong stocks (HK) - **🗣️ Natural Language**: Supports natural language queries without specifying market and date - **📈 Real-Time Decision Making**: Trading suggestions based on the latest data - **🌐 Web Frontend**: Complete Streamlit Web interface, supporting real-time analysis and historical management ## 🏗️ System Architecture ### Agent Organization Structure ``` ┌─────────────────────────────────────────────────────────────┐ │ TradingAgents-MCPmode │ ├─────────────────────────────────────────────────────────────┤ │ 📊 Analysts Team (分析师团队) - Parallel Execution │ │ ├── CompanyOverviewAnalyst (公司概述分析师) │ │ ├── MarketAnalyst (市场分析师) ┐ │ │ ├── SentimentAnalyst (情绪分析师) │ Parallel Processing │ │ ├── NewsAnalyst (新闻分析师) │ 6 Analysts │ │ ├── FundamentalsAnalyst(基本面分析师) │ Executing Simultaneously │ │ ├── ShareholderAnalyst (股东分析师) │ │ │ └── ProductAnalyst (产品分析师) ┘ │ ├─────────────────────────────────────────────────────────────┤ │ 🔬 Researchers Team (研究员团队) │ │ ├── BullResearcher (看涨研究员) │ │ └── BearResearcher (看跌研究员) │ ├─────────────────────────────────────────────────────────────┤ │ 👔 Managers (管理层) │ │ ├── ResearchManager (研究经理) │ │ └── Trader (交易员) │ ├─────────────────────────────────────────────────────────────┤ │ ⚠️ Risk Management Team (风险管理团队) │ │ ├── AggressiveRiskAnalyst (激进风险分析师) │ │ ├── SafeRiskAnalyst (保守风险分析师) │ │ ├── NeutralRiskAnalyst (中性风险分析师) │ │ └── RiskManager (风险经理) │ └─────────────────────────────────────────────────────────────┘ ``` ### 🚀 Parallel Workflow ```mermaid graph TD %% User Input USER[👤 User Input<br/>user_query] %% Stage 0: Company Overview Analyst A0[🏢 Company Overview Analyst<br/>📥 Input: user_query<br/>📤 Output: company_details + company_overview_report] %% Stage 1: Analyst Parallel Node PARALLEL[⚡ Analyst Parallel Node<br/>📥 Input: user_query + company_details<br/>🔄 Concurrently execute 6 analysts<br/>📤 Output: All 6 analysis reports] %% Concurrently executed 6 analysts A1[🔍 Market Analyst] A2[😊 Sentiment Analyst] A3[📰 News Analyst] A4[📊 Fundamental Analyst] A5[👥 Shareholder Analyst] A6[🏭 Product Analyst] %% Stage 2: Researcher Debate B1[📈 Bullish Researcher<br/>📥 Input: user_query + All 7 analyst reports<br/>📤 Output: Bullish Argument + Debate History] B2[📉 Bearish Researcher<br/>📥 Input: user_query + All 7 analyst reports + Debate History<br/>📤 Output: Bearish Argument + Debate History] %% Stage 3: Management C1[🎯 Research Manager<br/>📥 Input: user_query + All 7 analyst reports + Complete Debate History<br/>📤 Output: investment_plan] C2[💰 Trader<br/>📥 Input: user_query + All 7 analyst reports + Debate History + investment_plan<br/>📤 Output: trader_investment_plan] %% Stage 4: Risk Management Team D1[🔥 Aggressive Risk Analyst<br/>📥 Input: All Information<br/>📤 Output: Aggressive Risk Perspective + Risk Debate History] D2[🛡️ Conservative Risk Analyst<br/>📥 Input: All Information + Risk Debate History<br/>📤 Output: Conservative Risk Perspective + Risk Debate History] D3[⚖️ Neutral Risk Analyst<br/>📥 Input: All Information + Risk Debate History<br/>📤 Output: Neutral Risk Perspective + Risk Debate History] D4[🎯 Risk Manager<br/>📥 Input: All Information + Complete Risk Debate History<br/>📤 Output: final_trade_decision] %% Main Flow Connections USER --> A0 A0 --> PARALLEL PARALLEL --> B1 B1 --> B2 B2 -.->|Debate Loop| B1 B1 --> C1 C1 --> C2 C2 --> D1 D1 -.->|Risk Debate Loop| D2 D2 -.->|Risk Debate Loop| D3 D3 -.->|Risk Debate Loop| D1 D1 --> D4 D2 --> D4 D3 --> D4 %% Internal Connections of Parallel Node (dashed lines indicate concurrency) PARALLEL -.-> A1 PARALLEL -.-> A2 PARALLEL -.-> A3 PARALLEL -.-> A4 PARALLEL -.-> A5 PARALLEL -.-> A6 %% Style Definitions style USER fill:#f9f9f9,stroke:#333,stroke-width:2px style A0 fill:#e1f5fe,stroke:#0277bd style PARALLEL fill:#fff3e0,stroke:#ff8f00,stroke-width:3px style A1 fill:#e8f5e8,stroke:#4caf50 style A2 fill:#e8f5e8,stroke:#4caf50 style A3 fill:#e8f5e8,stroke:#4caf50 style A4 fill:#e8f5e8,stroke:#4caf50 style A5 fill:#e8f5e8,stroke:#4caf50 style A6 fill:#e8f5e8,stroke:#4caf50 style B1 fill:#e3f2fd,stroke:#1976d2 style B2 fill:#e3f2fd,stroke:#1976d2 style C1 fill:#fff3e0,stroke:#ef6c00 style C2 fill:#fff3e0,stroke:#ef6c00 style D1 fill:#fce4ec,stroke:#c2185b style D2 fill:#fce4ec,stroke:#c2185b style D3 fill:#fce4ec,stroke:#c2185b style D4 fill:#fce4ec,stroke:#c2185b ``` ### ⚡ Advantages of Parallelization 1. **Significant Efficiency Improvement**: Six analysts executing concurrently, with total time close to that of the slowest analyst's execution time. 2. **Resource Optimization**: Fully utilizes system resources, avoiding serial waiting. 3. **Data Consistency**: Each analyst uses a deep copy of the state to avoid concurrent conflicts. 4. **Result Merging**: Intelligent merging of reports and execution histories from each analyst. #### Online Experience (Public Address) - Experience Address: `http://47.79.147.241:8501` - Description: Open to the public for free to experience and demonstrate features. - Note: Please use it responsibly and do not engage in any form of attacks, stress testing, or malicious access to avoid affecting service stability. <img width="1920" height="1080" alt="image" src="https://github.com/user-attachments/assets/c557e15a-de1e-4dbe-b619-b8f32a2a91c2" /> ## 🚀 Quick Start ### Environment Requirements - Python 3.8+ - Supported Operating Systems: Windows, macOS, Linux ### Installation Steps 1. **Clone the Project** ```bash git clone https://github.com/guangxiangdebizi/TradingAgents-MCPmode.git cd TradingAgents-MCPmode ``` 2. **Install Dependencies** ```bash pip install -r requirements.txt ``` 3. **Configure Environment Variables** ```bash cp env.example .env ``` # Edit the .env file to configure your API key and workflow parameters ``` 4. **Configure MCP Tool** ```bash # Edit the mcp_config.json file to configure the MCP Server ``` ### 🌐 Web Frontend Usage **Recommended Usage** - Interact through the web interface: ```bash streamlit run web_app.py ``` Then access in your browser: `http://localhost:8501` <img width="1920" height="1080" alt="image" src="https://github.com/user-attachments/assets/280599ee-c00a-4c8c-b61f-c786e43cc6d4" /> #### Web Frontend Features - **🏠 Home**: System overview and quick navigation - **⚙️ System Configuration**: Online editing of .env and MCP configuration - **🔍 Real-time Analysis**: Input queries and monitor analysis progress in real-time - **🤖 Agent Control**: Dynamically enable/disable specific agents with a checkbox interface organized by team - **🌀 Debate Configuration**: Adjust investment and risk debate rounds in real-time for precise control over analysis depth - **📊 Agent Page**: View detailed analysis results of each agent by team - **📚 Historical Reports**: Manage historical sessions, supporting export to Markdown, PDF, and DOCX #### 🎛️ Frontend Agent Control Features **Agent Enable Control**: - ✅ **Group Management**: Organized by four main teams: Analysts, Researchers, Management, and Risk Management - ✅ **Real-time Selection**: Checkbox interface supporting quick operations for select all/deselect all - ✅ **Dynamic Count**: Real-time display of the number of enabled agents (e.g., 12/15 enabled) - ✅ **Smart Skip**: Disabled agents will be automatically skipped, ensuring workflow continuity **Debate Round Configuration**: - 🔄 **Investment Debate**: Slider to set bullish/bearish researcher debate rounds (each round = one speech from each side) - ⚖️ **Risk Debate**: Slider to set aggressive/conservative/neutral risk analyst debate rounds (each round = one speech from each of the three parties) - 📊 **Real-time Preview**: Displays current settings and expected number of speeches - 🚀 **Automatic Application**: Automatically applies frontend settings when analysis begins, no manual confirmation required ### 💻 Command Line Usage If you prefer using the command line: ```bash python main.py -c "Analyze Apple Inc. stock" ``` #### Interactive Mode ```bash python main.py ``` Then follow the prompts to enter your query, for example: - "Analyze Tesla stock" - "Give me an analysis of 600036 China Merchants Bank" - "Analyze the investment value of NVDA" ## 📊 Data Flow Mechanism ### 🎯 Key Design Features 1. **company_details Placeholder System** - 🏢 **Company Overview Analyst** first obtains the basic information about the company - 📥 **Only passed to the Analyst Team**: 6 professional analysts can access accurate company background - 🚫 **Not passed to subsequent agents**: Researchers and later agents focus on the comprehensive evaluation of the analysis report 2. **Information Accumulation Effect** - **Phase 0**: User query only - **Phase 1**: User query + Company details (parallel processing) - **Phase 2**: User query + All 7 analyst reports - **Phase 3**: User query + Analyst reports + Debate history + Investment decisions - **Phase 4**: User query + All information + Risk perspectives 3. **Dual Debate Mechanism** - **Investment Debate**: Bullish ↔ Bearish analysts engage in a cyclical debate - **Risk Debate**: Aggressive ↔ Conservative ↔ Neutral risk analysts engage in a cyclical debate ### 💡 Practical Application Example **User Input**: "Analyze Tesla stock" 1. **🏢 Company Overview Analyst**: Gather basic information about Tesla Inc., TSLA, NASDAQ, electric vehicle industry, etc. 2. **⚡ Analyst Parallel Nodes**: 6 analysts conduct professional analysis simultaneously based on accurate information about "Tesla Inc." 3. **📈 Bullish Researcher**: Integrate all 7 professional analysis reports to construct a bullish investment thesis for Tesla. 4. **🎯 Risk Manager**: Make the final risk decision on Tesla investment based on all information and debate results. This design ensures **accurate information transmission** and **gradual refinement**, allowing each agent to leverage their expertise in the most suitable information environment! ## 🎯 Usage Example ### 📱 Frontend Operation Example **Scenario**: Quickly analyze Apple Inc. stock, enabling only core analysts and setting up a simplified debate. 1. **Start the System**: ```bash streamlit run web_app.py ``` 2. **Configure Agents** (expand in "🤖 Enable Agents for This Round"): - ✅ Analyst Team: Select All (7 agents) - ✅ Researcher Team: Select only bullish researchers - ❌ Management: Deselect All (quick analysis) - ❌ Risk Management: Deselect All (quick analysis) - 📊 Display: Enabled 8/15 3. **Set Debate Rounds** (expand in "🌀 Debate Rounds Settings"): - 🔄 Investment Debate Rounds: 1 round (only bullish researchers speak once) - ⚖️ Risk Debate Rounds: 0 rounds (skip risk debate) 4. **Execute Analysis**: - Input: "Analyze the investment value of Apple Inc. stock" - Click "🚀 Start Analysis" - The system automatically applies the frontend settings and begins execution. 5. **Expected Results**: - ⚡ Quick Execution: Only 8 agents involved, significantly reducing analysis time - 📊 Core Report: Obtain 7 professional analyst reports - 🎯 Investment Advice: Bullish researchers provide investment advice based on analyst reports. ### 🎛️ Advantages of Dynamic Configuration - **🚀 Fast Experience**: Disable non-core agents to obtain analysis results in 1-2 minutes - **🔬 In-Depth Analysis**: Enable all agents + multi-round debates for comprehensive decision support - **🎯 Customization**: Flexibly choose agent combinations based on analysis needs - **💡 Real-Time Adjustment**: Adjust parameters in real-time on the front end without restarting the service ## 🛠️ Configuration Instructions ### Environment Variable Configuration Configure the following parameters in the `.env` file: ```env # Large Model Configuration OPENAI_API_KEY=your_openai_api_key OPENAI_BASE_URL=https://api.openai.com/v1 MODEL_NAME=gpt-4 # Workflow Configuration MAX_DEBATE_ROUNDS=1 # Default rounds for investment debate (can be dynamically adjusted in the frontend) MAX_RISK_DEBATE_ROUNDS=1 # Default rounds for risk debate (can be dynamically adjusted in the frontend) DEBUG_MODE=true # Debug mode, displays detailed logs VERBOSE_LOGGING=true # Detailed log output # Intelligent Agent MCP Permission Configuration # Analyst Team - It is recommended to enable MCP for real-time data COMPANY_OVERVIEW_ANALYST_MCP=true MARKET_ANALYST_MCP=true SENTIMENT_ANALYST_MCP=true NEWS_ANALYST_MCP=true FUNDAMENTALS_ANALYST_MCP=true SHAREHOLDER_ANALYST_MCP=true PRODUCT_ANALYST_MCP=true # Researcher/Management/Risk Team - Suggest to Disable MCP Focused Comprehensive Analysis BULL_RESEARCHER_MCP=false BEAR_RESEARCHER_MCP=false RESEARCH_MANAGER_MCP=false TRADER_MCP=false AGGRESSIVE_RISK_ANALYST_MCP=false SAFE_RISK_ANALYST_MCP=false NEUTRAL_RISK_ANALYST_MCP=false RISK_MANAGER_MCP=false # Task Concurrency Limit MAX_CONCURRENT_ANALYSIS=2 # Number of analysis tasks running concurrently ``` > **💡 Tip**: The environment variables `MAX_DEBATE_ROUNDS` and `MAX_RISK_DEBATE_ROUNDS` are only default values and can be adjusted in real-time on the front end during actual operation. It is recommended to set them to 1 for a quick experience, and if in-depth analysis is needed, you can increase the rounds on the front end. ### MCP Tool Configuration Configure the MCP server in the `mcp_config.json` file: ```json { "mcpServers": { "finance-mcp": { "disabled": false, "timeout": 600, "transport": "streamable_http", "url": "https://finvestai.top/mcp", "headers": { "X-Tushare-Token": "Your Tushare token" } } } } ``` **Configuration Description**: - `disabled`: Whether to disable the MCP server (false means enabled) - `timeout`: Request timeout duration (in seconds), default is 600 seconds - `transport`: Transport protocol, use `streamable_http` - `url`: MCP server address `https://finvestai.top/mcp` - `X-Tushare-Token`: Please replace with your actual Tushare API Key > **💡 Tip**: After modifying the configuration, please restart the application or hot reload the MCP client to apply the changes. ## 📈 Performance Optimization ### Advantages of Parallel Processing - **Original Serial Architecture**: 6 analysts execute sequentially, total time = Σ(time taken by each analyst) - **New Parallel Architecture**: 6 analysts execute concurrently, total time ≈ max(time taken by each analyst) - **Performance Improvement**: Theoretically, efficiency can be improved by 5-6 times; actual improvement depends on the time distribution of each analyst ### Resource Management - Use deep copy to avoid state contention - Smartly merge execution history and tool invocation records - Optimize memory usage to avoid state redundancy ## 🤝 Contribution Guidelines We welcome you to submit Issues and Pull Requests to improve the project! ## 📄 License This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details. --- **TradingAgents-MCPmode** - Let AI agent teams safeguard your investment decisions! 🚀📈 ## 📬 Contact Information - Email: `guangxiangdebizi@gmail.com` - LinkedIn: [Xingyu Chen](https://www.linkedin.com/in/xingyu-chen-b5b3b0313/) - Facebook: [Personal Page](https://www.facebook.com/profile.php?id=100072282093932) - Bilibili: [Personal Space](https://space.bilibili.com/51239486?spm_id_from=333.1007.0.0)