Available
—
models loaded
Total local
—
in pool
llama.cpp
—
models
Ollama
—
models
Disabled
—
excluded from routing
Local models

Model pool

Load/unload llama.cpp models. Assign surfaces. Ollama manages its own lifecycle automatically.

Runtime events

Live trace

[STATE] Loading…
Reasoning layers

Per-model KRL v3 — local models

Shared with Settings. Set independently per surface — Home, LoreCore, Pipeline.