NetworkOps_Platform

NetworkOps_Platform is an AI-powered network automation MCP server that provides a unified dashboard and a rich set of tools for managing multi-vendor networks. The platform exposes a suite of MCP tools (health checks, device discovery, status monitoring, and configuration validation) that let operators interact with Cisco, Juniper, Nokia, Arista, and Linux devices through natural language prompts or scripted commands. Use the dashboard to visualize topology, monitor device health, run automated checks, and perform common operations such as listing devices, checking connectivity, and validating network state. Tools can be invoked from the MCP client to perform targeted actions, get structured outputs, and receive source citations for results when using the AI-assisted querying mode.

Prerequisites:

• Python 3.11+ installed on the host
• Git installed
• Optional: Node.js and npm if you wish to build the dashboard front-end locally

1. Clone the repository: git clone https://github.com/E-Conners-Lab/NetworkOps_Platform.git cd NetworkOps_Platform

2. Set up a Python virtual environment and install dependencies: python3 -m venv .venv source .venv/bin/activate pip install -r requirements.txt

3. Install the dashboard dependencies and build assets (optional if using prebuilt assets): cd dashboard npm install npm run build cd ..

4. Configure environment variables (example):

   • JWT_SECRET: A secret string for JWT authentication
   • DEMO_MODE: true to run in demo mode (no real devices)
   • DB_CONNECTION: Optional database connection string if using persistence

5. Run the MCP server (as per mcp_config): Ensure you are in the project root and the virtual env is active, then start the API server: python dashboard/api_server.py

6. Optional Docker-based run:

   • Follow the repository's docker-init.sh script to prepare and run containers: ./scripts/docker-init.sh
   • Then use docker compose up to start services

7. Access the UI: Open http://localhost:5001 (or the port configured in the environment) to view the MCP-enabled dashboard.

Notes and tips:

• The platform includes a large set of MCP tools (Health & Monitoring, Device Operations, etc.). Use the health_check and full_network_test tools to validate multi-vendor connectivity and routing health before making changes.
• In Demo Mode, many features simulate data and do not require real devices. Real device integrations require provisioning device IPs/credentials in .env and corresponding config files.
• If you enable ANTHROPIC_API_KEY or other AI-related keys for advanced RAG/chat capabilities, ensure those keys are securely managed and loaded in the environment where the server runs.
• For production, consider enabling TLS/HTTPS in front of the API server and securing JWT_SECRET.
• If you modify configuration for devices, keep a backup of devices.py and related config files. The MCP tools rely on a consistent inventory to produce accurate results.
• The npm/dashboard components, if used, should be built to match the Python backend version to avoid compatibility issues.
• Common issues often involve environment variables not being picked up, or network access blocking required ports. Verify that port 5001 (or your configured port) is open and reachable.