AI Assistant

AI Assistant (optional)

Boltarr can use an LLM to analyze hosts and your network, and to power a chat assistant that has full context of what Boltarr knows. It's entirely optional — leave the provider on none and everything else works.

Providers

Provider	Use for	Needs
Ollama	Local / self-hosted models (private, free)	Base URL (e.g. http://ollama-host:11434), model
OpenAI-compatible	OpenAI, LM Studio, Groq, Together, …	Base URL, API key, model
Anthropic	Claude API	API key, model

Setup

Settings → 🤖 AI (or via environment variables — see Configuration):

• Provider, Base URL (Ollama / OpenAI-compatible), API Key (OpenAI / Anthropic), Default Model.
• Timeout / Long timeout — the long one covers big jobs like a full network analysis.
• ⚡ Test AI to confirm the connection.

For Ollama, pull a model first (e.g. ollama pull llama3.2) and set it as the default.

What it does

• Per-host analysis — from a host's detail panel, get an LLM summary of what the device likely is and any notable open ports.
• Network-wide analysis — a big-picture look across everything Boltarr has discovered.
• Chat — ask questions about your network; the assistant has context from your hosts, services, and scans.

Privacy note

With Ollama (local), nothing leaves your network. With OpenAI/Anthropic, the host/network data included in a request goes to that provider — use a local model if that matters to you.