AI Features

Conversational AI assistance for network operations with automatic credential sanitization, terminal context injection, and persistent history.

AI command autocomplete (inline ghost text accepted with Tab) and next-step suggestions based on your last command and its output.

Autonomous AI agents that monitor, triage, and respond to network events using the ReAct pattern with tool approval workflows.

How a NetStacks agent delegates work to specialist or ephemeral child agents, the workflow graph, parking/await, and the fail-closed guardrails that bound a delegation tree.

Mandatory redaction before any text reaches an LLM: 19 built-in patterns, optional network redaction, custom regex, and a case-insensitive allowlist.

Build a vector-searchable repository of organizational knowledge that powers AI Chat and NOC Agents with RAG-grounded answers.

Crawl documentation sites with a same-origin BFS crawler and ingest the extracted content into the NetStacks Knowledge Base for vector search.

Connect NetStacks to OpenAI, Anthropic, Ollama, OpenRouter, LiteLLM, or any OpenAI-compatible LLM. BYOK keys, models, token budgets, and self-hosting.

Connect Model Context Protocol servers to extend NetStacks AI agents and chat with external tools over stdio or SSE transports.

Built-in vendor and domain knowledge packs ground the assistant for Cisco, Juniper, Arista, routing, data center, MPLS, security, and wireless — loaded by profile weight within a budget.

NetStacks AI has two agent types — Auto Pilot and Overlord — combined with three permission modes (Ask, Auto, Go Nuts) that govern how tool calls execute.

How NetStacks composes the AI system prompt from personality segments and weighted vendor/domain knowledge packs per feature, the character budget, and where to configure it.

Save reusable AI prompt templates in NetStacks, run them from the status bar with live session context, and share them across your org.

Saved one-click API calls bound to an API Resource: variable substitution in path, headers, and body, with JSON path extraction of the result.