A per-model table says whether a model has a thinking switch and which effort levels;
the composer shows a grey effort slider and a brain switch only when the model has them.
llama.cpp and vLLM get thinking as enable_thinking; the provider JSON tables ship in the package.
The picker groups models by where they run, with search and a way into Pick a model.
A ⚙ beside the model sets thinking or effort for this session only (new session route).
GET /v1/system: processor, graphics, memory, storage, the largest model that runs well.
GET /v1/providers/ollama/models: size, tools, thinking, context and fit per model, read from Ollama instead of guessed; only tool-capable models join the picker.
model_config.json holds context size, longest reply, thinking, sampling, auto-compact and the default model; the app's API and the CLI read it.
recommended_models.json carries the makers' own values per model, with sources; docs/model-settings.md explains both.