Inference Flow¶ Inference-oriented flows load user input, invoke an installed model or semantic library, and render Markdown. Keep model selection, hardware expectations, and cache locations explicit in deployment docs.