Lead the README with the graph, not the animation

Swaps the animated loop for a screenshot of the consistency grid actually
running: one Load feeding three Apply nodes, three unrelated subjects, three
renders that hold the same style. The animation only ever showed outputs — this
shows the mechanism and the result together, which is the thing worth proving.

Keeps assets/workflow.webp: it is the source the five template thumbnails were
cut from, now noted in .comfyignore so it does not read as dead weight.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

.comfyignore

@@ -16,10 +16,14 @@ tests/
 scripts/
 .github/
 
-# README artwork. The README references it by absolute URL, so GitHub and the
-# Registry both render it without the bytes riding along into every install.
-# workflows/*.webp are NOT excluded — those are the template thumbnails, and
-# ComfyUI reads them from disk next to each workflow JSON.
+# Artwork. consistency-grid.webp and gallery.webp are referenced by the README
+# via absolute URL, so GitHub and the Registry render them without the bytes
+# riding along into every install. workflow.webp is the animated source the
+# five workflows/*.jpg thumbnails were cut from — keep it, or regenerating them
+# means re-rendering the frames.
+#
+# workflows/*.jpg are NOT excluded: those are the thumbnails, and ComfyUI reads
+# them from disk next to each workflow JSON. They must ship.
 assets/
 
 # Tooling caches and editor droppings, in case a publish runs from a working

README.md

@@ -1,12 +1,14 @@
 # StyleRef nodes for ComfyUI
 
-<img src="https://raw.githubusercontent.com/StyleRef/styleref-comfyui/main/assets/workflow.webp" alt="One StyleRef style applied to seven different subjects" width="420" align="right">
-
 Define a creative style once on [StyleRef](https://styleref.io) — colors,
 lighting, mood, composition, guardrails — then load it as a node and reuse it
 across your workflows.
 
-Every frame on the right is the same style, a different subject.
+![One style, three unrelated subjects, rendered from a single StyleRef Load node](https://raw.githubusercontent.com/StyleRef/styleref-comfyui/main/assets/consistency-grid.webp)
+
+One **StyleRef Load** feeding three **Apply** nodes with three unrelated
+subjects — a rooftop at dusk, a record shop, a toy unicorn. The style holds;
+only the subject changes. That is the whole idea.
 
 The style is compiled server-side for whichever model you point it at, so the
 same style produces prompt text phrased appropriately for each one.