AI Assistant
Ask your infrastructure questions in plain language. The AI Assistant connects to OpenAI, Anthropic, or a local Ollama model and can call a permission-gated set of tools to look up devices, alerts, and VLANs, with governance applied before anything leaves your network.
Multi-provider LLM assistant for OpenAI, Anthropic, and Ollama over direct httpx. It runs an agentic loop over a permission-gated tool registry, with governance from a global kill-switch through per-org policy to PII redaction before any cloud call, and an audit trail on every invocation.

AI Assistant capabilities
Multi-provider LLM
Bring your own model: OpenAI, Anthropic, or a self-hosted Ollama instance, called directly over httpx.
Permission-gated tools
The assistant runs an agentic loop over a registry of typed tools. Each tool is gated by the same permissions a human would need, so it can never do more than the caller could.
Layered governance
A global kill-switch, per-org policy, and PII redaction all run before any cloud call, so you control exactly what a model can see.
Bounded agentic loop
Tool-calling is capped at a small number of iterations per request, so a runaway loop cannot run away.
Audit trails
Every invocation is recorded, so you can review what was asked and what the assistant did.
Grounded in your data
Suggested prompts such as "list all offline devices" or "check the health of switch-core-01" run against your live inventory.
Providers: OpenAI, Anthropic, and self-hosted Ollama, called directly over httpx.
- Beta. The assistant stays disabled until you configure a provider, and nothing is sent to any model provider unless you enable it.
Explore other modules
All modules →See it running in your browser
Explore the full FreeSDN dashboard with realistic sample data, no signup, no backend. Then install it in minutes.