LangChain Agent MCP Server

# Getting Started This guide will help you set up and run the LangChain Agent MCP Server locally. ## Prerequisites - Python 3.11 or higher - OpenAI API key - Git (optional, for cloning) ## Installation ### Step 1: Clone the Repository ```bash git clone https://github.com/mcpmessenger/LangchainMCP.git cd LangchainMCP ``` ### Step 2: Install Dependencies **Windows:** ```powershell py -m pip install -r requirements.txt ``` **Linux/Mac:** ```bash pip install -r requirements.txt ``` ### Step 3: Configure Environment Create a `.env` file in the project root: ```env OPENAI_API_KEY=your-openai-api-key-here OPENAI_MODEL=gpt-4o-mini PORT=8000 HOST=0.0.0.0 ``` ### Step 4: Run the Server **Windows:** ```powershell py run_server.py ``` **Linux/Mac:** ```bash python run_server.py ``` You should see: ``` Loaded environment variables from .env Starting LangChain Agent MCP Server on 0.0.0.0:8000 API Documentation: http://0.0.0.0:8000/docs ``` ## Verify Installation ### Test Health Endpoint ```powershell # Windows PowerShell Invoke-WebRequest -Uri "http://localhost:8000/health" # Linux/Mac curl http://localhost:8000/health ``` Expected response: ```json {"status":"healthy"} ``` ### Test MCP Manifest ```powershell Invoke-WebRequest -Uri "http://localhost:8000/mcp/manifest" ``` ### Test Agent Invocation ```powershell $body = @{ tool = "agent_executor" arguments = @{ query = "What is 2+2?" } } | ConvertTo-Json Invoke-WebRequest -Uri "http://localhost:8000/mcp/invoke" ` -Method POST ` -ContentType "application/json" ` -Body $body ``` ## Next Steps - [Deploy to Google Cloud Run](deployment.md) - [Explore Examples](examples.md) - [Read API Documentation](api-reference.md)