Network AI Assistant

A Streamlit-based AI agent for network assistance with OSPF/IS-IS capabilities. Works with Topolograph MCP.
Interact naturally with OSPF/IS-IS protocols, for instance you can ask:

• Which graphs are currently connected?
• What nodes are in the latest OSPF area 0 graph?
• Which networks are assigned to a specific host?
• What is the route between two IP addresses?
• What happened to the links after topology changes?

Quick Start

Prerequisites

0. Register on OpenAI and generate API token:
   • Go to https://platform.openai.com/
   • Create an account and generate an API token
   • Note: For repeating the demo setup, you'll need approximately less than $1 in your OpenAI balance
   • Note: Local LLM setup using vLLM is under development

Using Docker Compose

Current setup uses public OpenAI models, so it means that if you use local MCP connection, it's needed to expose local MCP to the OpenAI.

1. Establish tunnel to make your local MCP publicly available

Cloudflare Tunnel

1. Run with Cloudflare tunnel:

   docker-compose --profile cloudflare up --build

2. Wait for tunnel establishment:

   • The Cloudflare tunnel will create a secure connection
   • The public tunnel URL will be displayed in the logs
   • Look for output like: https://your-tunnel-url.trycloudflare.com

Ngrok (Free alternative)

If you prefer ngrok (requires free registration at https://ngrok.com):

1. Install ngrok:

   # Download from https://ngrok.com/download
   # Or using package manager
   sudo apt install ngrok  # Ubuntu/Debian

2. In a separate terminal, establish ngrok tunnel:

   # For the Streamlit web UI (port 8501)
   ngrok http 8501
   
   # For Topolograph's MCP server (port 8080)
   # Topolograph's MCP server is published via Nginx, which is available via 8080 `TOPOLOGRAPH_PORT` variable in `env` file
   ngrok http 8080

2. Copy URL into .env

Copy the tunnel URL (from Cloudflare or ngrok) into your .env file:

# Add your tunnel URL here
MCP_SERVER_URL=https://your-tunnel-url.trycloudflare.com
# or
MCP_SERVER_URL=https://abc123.ngrok.io

3. Start the application

1. Run the application:

   docker-compose up --build

2. Access the web UI:

   • Local: http://localhost:8501

Demo Video

Watch the Network AI Assistant in action: YouTube Demo

Demo Setup

To repeat the same demo on your host:

1. Clone and setup Topolograph with OSPF lab:

   git clone https://github.com/Vadims06/topolograph-docker
   cd topolograph-docker
   sudo ./install.sh

   This will setup all required packages and start the Topolograph service with MCP protocol and OSPF local lab.

2. Start the Network AI Assistant:

   # In this repository
   docker-compose up --build

Screenshots

Main Application Interface

OSPF Network Analysis

Network State After Link Changes

Development

The application uses Streamlit and includes MCP (Model Context Protocol) functions for network analysis.

Environment Variables

Copy from .env.template and configure:

cp .env.template .env
# Edit .env with your configuration

Local Development

pip install -r requirements.txt
streamlit run app.py

Online Resources. Contacts

• Telegram group: https://t.me/topolograph
• Main site: https://topolograph.com
• Docker version of site: https://github.com/Vadims06/topolograph-docker
• MCP: https://github.com/Vadims06/topolograph-mcp-server