AI Chat

Interactive AI assistance for network operations.

Overview

AI Chat provides conversational assistance for network operations. Ask questions, get command suggestions, and troubleshoot issues.

Natural language interface
Context-aware responses based on your session
Knowledge base integration
Multi-turn conversations

Using AI Chat

Press Cmd/Ctrl+Shift+A to open AI Chat
Type your question or request
AI responds with explanations or commands
Click suggested commands to insert them

Example Questions

"How do I check BGP neighbor status?"
"What's wrong with this OSPF configuration?"
"Generate an ACL to block traffic from 10.0.0.0/8"
"Explain this error message"

Context Awareness

AI Chat understands your current context:

Device Type - Provides platform-specific commands
Recent Output - Analyzes terminal output for troubleshooting
Selected Text - Explains highlighted configuration
Conversation History - Remembers previous messages

Tip

Select terminal output before asking a question to give AI context for analysis.

Data Privacy & Security

NetStacks automatically protects sensitive data before sending any information to AI providers. A built-in sanitization layer detects and redacts credentials, secrets, and other sensitive information from your conversations.

What Gets Protected

Category	Examples
Device Credentials	Passwords, enable secrets, username/password combinations
SNMP Communities	Community strings (v1/v2c), SNMPv3 auth and priv keys
Authentication Keys	TACACS+ keys, RADIUS secrets, API tokens
Routing Protocol Secrets	BGP MD5 passwords, OSPF authentication, EIGRP keys
VPN & Tunnel Keys	IPSec pre-shared keys, GRE tunnel keys, WireGuard keys
Certificates & Private Keys	RSA/DSA/ECDSA private keys, certificate data

How It Works

Pattern Detection - When you send a message or share terminal output, NetStacks scans for known credential patterns across all major network vendors (Cisco, Juniper, Arista, Palo Alto, and more).
Automatic Redaction - Detected secrets are replaced with safe placeholders like [REDACTED:password] before the data leaves your environment.
Audit Logging - All redactions are logged so you can review what was sanitized and verify protection is working correctly.

Example

If you share configuration containing:

username admin password 0 MySecretPass123
snmp-server community NetworkRO RO

The AI receives:

username admin password 0 [REDACTED:password]
snmp-server community [REDACTED:snmp-community] RO

Tip

For maximum privacy, use the self-hosted Ollama integration. When configured with a local LLM, your data never leaves your network. See the LLM Configuration page for setup instructions.

Enterprise Deployment

In Enterprise Edition, all AI processing is handled centrally by the Controller. Terminal clients act as thin clients and never communicate directly with AI providers. This architecture ensures consistent security policies and centralized audit logging across your entire organization.