NetworkBot – AI‑powered network monitoring

Overview

NetworkBot is a web-based AI assistant for network monitoring. It uses a large‑language model (LLM) to answer questions about your infrastructure, with optional integration to UniFi Network and UniFi Site Manager.

• Chat with the AI about status, logs, and diagnostics in a Matrix-style web UI.
• Fetch current metrics from your monitoring system or paste log snippets for analysis.
• Configure LLM (OpenAI or Ollama), monitoring integrations, and server settings from the same interface.

NetworkBot is Node.js + Express and supports OpenAI or Ollama as the LLM provider.

Features

Feature	How it works
Web Chat	Ask questions in the browser; the AI uses monitoring data when available and returns structured, human-readable answers.
Web Configuration	Configure LLM provider, API keys, models, UniFi Network/Site Manager, port, and auth—all from the UI.
UniFi integrations	Optional: pull devices and clients from UniFi Network; sites and devices from UniFi Site Manager (cloud).
Structured AI output	Status summary tables, error checks, conclusion, and next steps for easy scanning.
Docker	Docker Compose stack for deployment.

📦 Installation

1. Clone the repo:

   git clone https://github.com/RantsRoamer/networkbot/networkbot.git
   cd networkbot

2. Create a .env file (optional; you can configure everything in the web UI):

   # LLM Provider (choose one)
   LLM_PROVIDER=openai  # or 'ollama'
   
   # For OpenAI
   OPENAI_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
   OPENAI_MODEL=gpt-4o-mini
   
   # For Ollama (if LLM_PROVIDER=ollama)
   # OLLAMA_BASE_URL=http://localhost:11434
   # OLLAMA_MODEL=llama2
   
   PORT=3000

3. Install dependencies:

   npm install

4. Run:

   npm start

   The app runs a single server on http://localhost:3000 (or the port you set in config).
5. Open the web interface:
   • Go to http://localhost:3000
   • Log in (default: admin / admin — change this in Configuration!)
   • Use Chat to talk to the AI and Config to set LLM, monitoring, and server options.

🌐 Web Interface

Accessing the Web Interface

1. Start NetworkBot: npm start
2. Open your browser to http://localhost:3000
3. Log in with your credentials (default: admin / admin)

Features

• ✅ Switch LLM Providers: Toggle between OpenAI and Ollama with one click
• ✅ Configure API Keys: Securely store and update API keys
• ✅ Model Selection: Choose different models for each provider
• ✅ Monitoring: Configure UniFi Network and UniFi Site Manager from the web UI—add/remove controllers, test connections
• ✅ Test Connections: Verify Ollama and each monitoring integration before saving
• ✅ Server Settings: Adjust port and log levels
• ✅ Real-time Updates: Changes take effect immediately (some may require restart)

Security

• The web interface is protected with HTTP Basic Authentication
• Default credentials: admin / admin - CHANGE THIS IN PRODUCTION!
• Set custom credentials via environment variables:

  WEB_AUTH_USERNAME=your_username
  WEB_AUTH_PASSWORD=your_secure_password

• Or configure through the web interface itself (stored in config.json)

Configuration Storage

• Settings are saved to config.json in the project root
• Environment variables take precedence over config file
• The web interface updates config.json directly

🤖 Ollama Setup

NetworkBot supports Ollama for local LLM inference. This is perfect for privacy-sensitive environments or when you want to avoid API costs.

1. Install Ollama:

   • Visit https://ollama.ai and download for your platform
   • Or use Docker: docker run -d -p 11434:11434 ollama/ollama

2. Pull a model:

   ollama pull llama2
   # or
   ollama pull mistral
   # or any other supported model

3. Configure NetworkBot:

   LLM_PROVIDER=ollama
   OLLAMA_BASE_URL=http://localhost:11434
   OLLAMA_MODEL=llama2

4. Verify Ollama is running:

   curl http://localhost:11434/api/tags

Note: If running NetworkBot in Docker and Ollama on the host, use OLLAMA_BASE_URL=http://host.docker.internal:11434 or configure Docker networking.

📡 UniFi Network monitoring

NetworkBot can pull metrics from UniFi Network (devices, clients) using the official UniFi Network API.

• API reference: UniFi Network API – Getting started
• API key: Create in Network → Control Plane → Integrations (or Controller Settings → API Access on older versions).
• Base URL: Use your controller URL (e.g. https://10.69.69.1 or https://unifi.example.com). For UDM Pro / UniFi OS, the app uses /proxy/network or /unifi-api/network automatically.
• Authentication uses the X-API-Key header when supported; local controllers fall back to session login with the same key.

Configure one or more UniFi Network controllers in the web UI under Configuration → Monitoring → UniFi Network.

☁️ UniFi Site Manager (cloud)

NetworkBot can pull sites and devices from the cloud UniFi Site Manager API for a single UI account.

• API reference: UniFi Site Manager API – Getting started
• Base URL: https://api.ui.com (default)
• Authentication: X-API-Key header. Create a key in your UI account: Settings → API Keys (EA) or API section (GA).
• Rate limits: EA 100 req/min; v1 stable 10,000 req/min (read-only).

Enable in the web UI under Configuration → Monitoring → UniFi Site Manager (cloud) (checkbox + API key + optional base URL), or in config.json → monitoring.siteManager. Test with Test Site Manager or POST /api/monitoring/test-site-manager.

🚀 Usage

• Open the Chat tab and type a question (e.g. “Summarize UniFi device status” or “What do these logs indicate?”).
• The AI uses monitoring data when configured and returns a structured reply (status summary, error check, conclusion, next steps).
• Use the Configuration tab to set the LLM provider, API keys, UniFi integrations, and server port/auth.

🔧 Architecture

• Express serves the web UI and API (chat, config, monitoring tests).
• LLM queries go to OpenAI or Ollama based on configuration.
• Monitoring (optional) pulls data from UniFi Network and Site Manager and injects it into the AI context.

📁 Project Structure

networkbot/
├─ app.js             # Express server, API routes, chat & config
├─ public/            # Web UI (Matrix-style)
├─ utils/
│ ├─ llm.js          # LLM query helper (OpenAI / Ollama)
│ └─ logAnalyzer.js  # Config & monitoring (see config.js, monitoring.js)
├─ docker-compose.yml
├─ Dockerfile
├─ package.json
└─ README.md

Extending

Goal	Where to extend
Query Prometheus	utils/prometheus.js – add a function that runs curl or uses the official client and formats output for the LLM
Persist logs	Add a small SQLite database and utils/logger.js
Multi‑language LLM	Replace the OpenAIApi call with a LangChain provider

Contributing

1. Fork the repo.
2. Create a feature branch (git checkout -b feature/xxxx).
3. Add tests under __tests__.
4. Ensure npm test passes.
5. Open a pull request.

License

MIT – feel free to adapt and distribute.

---

Tip – If you want a ready‑loaded environment, use the Docker stack below.

🎉 Docker Compose

The included docker-compose.yml provides:

• NetworkBot service with health checks
• Optional Ollama service (commented out by default)
• Proper networking configuration

Build & run

docker compose up --build

The web interface will be available at http://localhost:3000.

Running with Ollama in Docker

Uncomment the ollama service in docker-compose.yml and set OLLAMA_BASE_URL=http://ollama:11434 in your .env file.

✨ Features & Improvements

• ✅ Web Configuration Interface: Beautiful UI for managing all settings
• ✅ Dual LLM Support: Works with both OpenAI and Ollama
• ✅ Robust Error Handling: Graceful error messages and logging
• ✅ Matrix-style UI: Dark theme with green/cyan terminal aesthetic
• ✅ Enhanced Log Analysis: Extracts errors, warnings, IPs, timestamps, and URLs
• ✅ Environment Validation: Checks required config on startup
• ✅ Health Checks: Docker health monitoring included
• ✅ Better Logging: Configurable log levels and structured output
• ✅ Configuration Management: JSON-based config with web UI and env var support