A save copies to a workflow card, a preset of a card with the same structure, or a
combo; a card's preset can join the open canvas's saves. The Workflows page lists
every group, so saves of a closed tab or a deleted file can be opened again, on a
new tab where the group goes on, or removed. A card keeps no key of a canvas's saves.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
AMB read models only from ordinary drop-downs, so rgthree's Power Lora Loader (rows are
{on, lora, strength} objects) was invisible: Model Check never reported its missing LoRAs.
Now its switched-on rows are Model Check boxes (lora_nodes.js powerLoraRowBoxes: ComfyUI's LoRA
list as choices, a fix changes only the row's LoRA; downloads put files in the same way), saved
fingerprints and a picture's model search read them, the workflow summary lists them, and saved
model references include them and Easy-Use LoRA stack rows (nodeModelValues /
_model_reference_values, mirrored). Only these widely used nodes; ComfyUI's loaders stay the
reference.
recipe_parser.js was over 600 lines: which LoRAs a node loads moves to recipe_loras.js and the
box readers to recipe_widgets.js.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
One module, prompt_boxes.js, now decides which boxes of a live node take a
prompt and whether each is positive or negative. A prompt box is a
multiline STRING input by the node's ComfyUI definition, not a widget
whose name looks right; its role is its own name (positive / negative) or
the wiring down to a sampler's positive / negative input, not the title.
- Material drops, Prompt Studio, the translator and recipe capture use it;
their name regexes and fallbacks ("any string widget") are gone.
- Prompt Studio's role detection read graph.links[id], which is a Map in
current ComfyUI, so it never followed a link; it uses the module now.
- Material writes call onWidgetChanged(name, ...) as ComfyUI expects.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keeps the planned All-in-One loader and cross-node prompt features but removes
the regressions they introduced:
- model references: one adapter table for verified layouts (easy a1111Loader,
easy fullLoader, Efficient Loader) shared by recipe_identity.js and
recipe_schema.py; native loaders keep their stored labels; no more guessing
categories from file names (a "...vae.safetensors" checkpoint was a VAE);
"None"/"Baked VAE" are never references
- recipe summary sync follows the extraction rules: only checkpoint/UNet
loaders or ckpt_name/unet_name set the base model (not an upscaler's
model_name), only sampler-like nodes set seed/steps/cfg, and LoRA edits
update the entry of the edited loader instead of falling back to loras[0]
- one LoRA entry per loader again (deduplication shifted entries)
- prompt injection never writes negative text into a positive slot or the
reverse; the block chooser follows the same rule
- material cards only accept cross-node drops that carry prompts; parameter
blocks stay same-type
- drop unused imports and the unused modelSpecs alias
Verified: regression cases added to tests/recipe_all_in_one_compatibility.mjs
and tests/material_prompt_cross_node.mjs (both fail on the previous commit),
new tests/test_recipe_model_references.py; all JS tests except the orphaned
material_parameter_cross_node.mjs and 95 Python tests pass.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeqEub5
