TuxNest ecosystem
Your compute.
A deliberate connection.
Use an authorised host and a supported model service when a local machine is not the right place for inference.
What remote inference means
SSH can protect the connection to a model service running on another host. That host receives prompts and selected context. A GPU is optional infrastructure on the host; SSH itself does not provide one.
You supply the host, the runtime and your authentication. Enter secrets only in supported application controls. Do not send private keys through chat, issue trackers or this website.
Before connecting
- Confirm you own the host or have explicit authorisation.
- Verify its identity and supported authentication method.
- Check that the intended model service is running.
- Check memory and compute capacity without oversubscribing it.
- Verify the selected route and the scope of data sent to it.