Combined MCP Server

# MCP LlamaIndex Chatbot Example A simple Streamlit chatbot that uses LlamaIndex with AWS Bedrock, connecting to the Combined MCP Server in HTTP mode. ## Prerequisites 1. Combined MCP Server running in HTTP mode: ```powershell $env:MCP_TRANSPORT="http" python -m combined_mcp_server.main ``` 2. AWS credentials configured with Bedrock access ## Installation ```bash cd examples pip install -r requirements.txt ``` ## Run ```bash streamlit run app.py ``` ## Features - Chat interface with AWS Bedrock Claude - MCP tool integration via HTTP/SSE transport - Knowledge base search using MCP tools - Conversation history ## Configuration Set these environment variables: - `AWS_REGION` - AWS region (default: us-east-1) - `BEDROCK_MODEL_ID` - Bedrock model (default: anthropic.claude-3-sonnet-20240229-v1:0) - `MCP_SERVER_URL` - MCP server URL (default: http://localhost:8000)