Inference Workers
Real-time cluster status of local office Ollama machines connected over outbound WebSocket.
Office Linux AI Host
LAN IP:
192.168.0.111:11434Tunnel Endpoint:wss://ai.techdrops.in/worker-ws
Heartbeat Rate:Every 15s
Firewall Security:No inbound ports exposed
AVAILABLE OLLAMA MODELS:
qwen2.5-coder:3bphi3.5:3.8bqwen2.5:3bllama3.2:latestllama3.2-vision:11bqwen2.5vl:3b
Worker Architecture & Security
- Outbound-Only Tunnel: The Linux worker initiates the connection to the VPS. Your office router requires zero port-forwarding.
- Dynamic Model Discovery: When you pull a new model via
ollama pull, the worker announces it to the gateway on the next heartbeat. - Automatic Reconnection: If office internet drops, the worker automatically retries with backoff and resumes handling developer jobs.