Introduction
GNETiX bridges your messaging platforms to AI agents and MCP tools
GNETiX is an AI-powered operations platform that connects Cisco Webex, Slack, Microsoft Teams, and webhooks to intelligent agents capable of querying and configuring your infrastructure in real-time.
How It Works
User (Webex/Slack/Teams/Voice) → GNETiX Director → MCP Tools → Your Infrastructure- A user sends a message through any supported platform
- The Director — GNETiX's cloud orchestration engine — receives the message
- The Director selects the best LLM, builds context from your agent souls and tools, and determines what actions to take
- Tool calls are dispatched to on-prem sub-agents via a secure outbound WebSocket relay
- Sub-agents execute tools against your infrastructure using the Model Context Protocol (MCP)
- Results flow back through the Director, which synthesizes a response and sends it to the user
No inbound firewall rules are ever required. All agent communication uses outbound WebSocket connections from your network.
Key Concepts
Director
Cloud-side LLM orchestration engine that processes messages, calls tools, and synthesizes responses.
Agents
On-prem containers that connect to GNETiX via WebSocket relay and execute MCP tools.
MCP Servers
FastMCP-based tool servers that expose your infrastructure capabilities (network devices, Kubernetes, monitoring, etc.).
Pipeline Monitor
Real-time visualization of every message flowing through the system — from ingestion to response.
Supported Platforms
| Platform | Status | Features |
|---|---|---|
| Cisco Webex | Production | Text, images, PDFs, voice calls |
| Microsoft Teams | Production | Text messaging |
| Slack | Production | Text messaging |
| Webhooks | Production | Inbound HTTP events from any system |
| Voice (SIP) | Production | Outbound calls via LiveKit + ElevenLabs |
LLM Providers
GNETiX supports multiple LLM providers through LiteLLM:
- Anthropic (Claude) — direct API
- OpenAI (GPT-4o) — direct API
- Amazon Bedrock — Claude via AWS (data stays in your VPC)
- Azure OpenAI — GPT models via Azure (enterprise compliance)
Each organization configures its own provider and API credentials. No data is shared between tenants.