Model Uplink Gives Your Local Model a Public Address

Field notes

The documented workflow starts with a local Ollama or vLLM server and ends with an OpenAI-compatible HTTPS endpoint, while a background service maintains the connection. The example uses `127.0.0.1:11434`, a named endpoint such as `home-gpu`, and a key shown once during setup. The page also describes relay routing through a Frankfurt region and lists editor, agent, RAG, and product workflows.