TuxNest · local-first AI engineering
Remote GPU
Connect a remote Ollama host over SSH so model compute can happen on hardware you control.
Built around control.
Minimal interface, explicit model choice, reviewable changes and evidence you can inspect.
01Prepare host
Run Ollama on the remote GPU machine.
02Secure access
Use SSH rather than exposing the runtime publicly.
03Tunnel endpoint
Forward the service to a loopback port.
04Connect TuxNest
Point the provider at the local end of the tunnel.
Local AI should feel local.
Your editor, your model runtime, your hardware and your approval remain first-class parts of the workflow.