Streamlit Chat App with MCP Support

A streamlit-based chat application with MCP (Model Context Protocol) server integration and multiple LLM support.

Current Status

Step 1: MCP Inspector for RBAC Testing ✅

• Streamlit app with proper MCP client implementation
• Connects directly to apstra_mcp.py HTTP transport (no separate API needed)
• Session-based authentication with Apstra credentials
• Tests all MCP protocol endpoints: initialize, list_tools, call_tool, list_prompts, list_resources
• Full MCP protocol compliance with proper RBAC authentication
• Validates that simple_http_api.py is not required

Quick Start

Using Docker (Recommended)

# Build and run with docker-compose
docker-compose up --build

# Access the app at http://localhost:8501

Local Development

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Run the app
streamlit run app.py

Testing MCP RBAC Authentication

1. Start the apstra-mcp-server with HTTP transport:

   cd ../apstra-mcp-server
   python apstra_mcp.py -t http -H 0.0.0.0 -p 8080

2. Open the MCP Inspector at http://localhost:8501

3. Test the authentication flow:

   • Login with your Apstra credentials
   • Run tests to verify proper MCP protocol and RBAC
   • Expected behavior:
     • Without login → Tools fail with auth error
     • With login → Tools work with user's identity
     • Session-based auth maintains state across calls

Key Findings

✅ simple_http_api.py is NOT required - The apstra_mcp.py server already provides:

• Full MCP protocol support with HTTP transport
• Session-based RBAC authentication
• All necessary endpoints for tool discovery and execution
• Proper user identity validation against Apstra

Next Steps

• Review and improve apstra-mcp-server HTTP API for RBAC (Not needed - use MCP server directly)
• Build proper chat interface with sidebar
• Add LLM connectors (OpenAI, Anthropic, Ollama)
• Implement proper MCP context flow to LLM