Send your first
local request.
By Alvin
In the prepared workshop Terminal, start the engine. Leave that window open.
python3 edge_workshop.py serve
Wait for Listening:. The model is loaded and the service is ready. Open a second Terminal, repeat the folder-drag cd and PATH steps, and run:
python3 edge_workshop.py ready python3 edge_workshop.py chat
Ready identifies the loaded model. chat sends one example question: “In two short sentences, explain what a Python program can do with a list of notes.” A reply confirms that your client reached the model and received text.
The service address 127.0.0.1:19437 means this computer, on the workshop’s chosen port. The client reads the model identifier from the local service. It uses an OpenAI-compatible message format; the request goes to Edge0 on this Mac and uses no hosted API key. Our client is a small one-request helper, not an interactive coding agent.
Use one model copy at a time. If ready cannot connect, wait for Listening: or check the server window. Press Control-C in that original window when you finish; the helper stops its owned engine processes.