Observed arrival · 2026-09-13
MetalMind puts an OpenAI-compatible local LLM server on your Mac
A native Apple Silicon app that runs open-weights language models locally and exposes them to editors through a localhost API.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
Field notes
MetalMind downloads model weights from Hugging Face in Apple's MLX format and runs them on the Mac's GPU and unified memory, without a cloud provider in the inference path. Its local server exposes /v1/chat/completions at 127.0.0.1:8080, supports SSE streaming, and queues concurrent requests rather than failing immediately. The free tier includes a request inspector with timings, payloads, and cURL export; the page also gives an example Qwen 2.5 7B 4-bit configuration.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
$PaidCommerce or pricing visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
◉HumanPersonal, local, civic, or handmade
One card from the complete issue