I'm working on local LLM inference. I want to save money on my Claude/Codex bill by having my Mac do it. I like Qwen for this.

I'm working on glass-slipper.cc (no telemetry; no login; just local offload from Claude to dumb model).

email: robertkarljr at the big google-provided email one.