FlywheelApp · full
Cinema + labs + WebGL. The labeled map, Hub canvas, field sketch, and every live engine. Use when this is the app that owns the GPU on the machine.Every feature, no graphics. Same engines, same API, same honesty grammar. No WebGL, no custom cursor, no Three.js. For when other apps already hold the GPU, or after learning loops when you only need the work.Terminal + one API out. Every live verb as a command and as HTTP. TUCE agents load skill.md and call /api/v1. No cinema overhead. Occupancy stays nominated | live-bound | unavailable.

Lab · 14 · Weights + KV + LoRA

Spoke budget

Add illustrative GiB for a hot set. Download-only is not running. Tokens/sec are not published.

NCM 1.19.0 · Hub 2.2.0 · invented_metrics = false

illustrative

Hot set

Add weights, KV, LoRA. Second LoRA fails closed.

This is the live spoke budget. Illustrative GiB. Accelerators stay unavailable. Tokens/sec are not published. Download-only is not running.

Add illustrative weight GiB for the selected class. Download-only is not running.Add illustrative KV GiB for context × batch 1. Not a serving SLA.Add illustrative LoRA GiB for the selected rank. A second LoRA compose fails closed.Reserve 4 GiB for the OS. Fit still gates load.Return the hot set to the seed rows.

kindlabelGiB
weights8B-class Q43.725
kvKV ctx 8192 × batch 11.000
loraLoRA rank 16 · 4 modules0.031
osOS reserve4.000

unavailable · single

8.757 GiB counted

Listed 8.757 GiB · hint unavailable · headroom unavailable

Accelerators stay unavailable. deviceMemory is a browser hint, not VRAM. No hint — stay unavailable rather than invent a fit.

Download the illustrative hot-set as JSON. Tokens/sec are not published.Copy this tab's URL with the budget rows in the querystring. Same origin only.Fit plannerIllustrative fit from this-tab memory, weight class, context, and LoRA rank. Accelerators stay unavailable.