Nodes/AAA Metadata System/LoRA Gallery with Edit v03
ComfyUI Node

LoRA Gallery with Edit v03

A LoRA Gallery You Can Actually Edit From

By EricRollei·Created 10 months ago·Updated 9 months ago· 13
LoRA Gallery with Edit v03
    • gallery_html
    • selected_index
    • selected_lora_info
    • edit_feedback
    lora_list
    selected_index1
    edit_modefalse
    gallery_sizemedium
    show_architecturetrue
    show_categorytrue
    show_ratingstrue
    show_triggerstrue
    filter_architectureAny
    filter_categoryAny
    min_rating0
    has_images_onlyfalse
    has_triggers_onlyfalse
    search_text

    The plain Gallery Display node shows you your LoRAs; LoRA Gallery with Edit v03 is the version that lets you do something about them. It's the same HTML-card browse interface, plus quick-edit mode, rating stars, trigger-word tooltips, and a set of filters - so you can look at your collection, spot the LoRA that's mis-categorized or under-rated, and fix it without leaving the gallery. It's the "action" sibling to a pure display node, and it's the one you'd actually build a daily driver around.

    Everything from the display version is here - lora_list, selected_index (1-based), gallery_size, show_architecture, show_category, and the architecture/category filters - with additions that make it an editing surface: edit_mode flips cards into an editable state, show_ratings displays the star ratings from the family's feedback nodes, and show_triggers puts trigger words into card tooltips. The new filters are where you'll spend your time: min_rating (0–5, 0 shows all), has_images_only (show only LoRAs with associated images), has_triggers_only (only ones with trigger words recorded), and search_text for a quick name/trigger search.

    Outputs are gallery_html, selected_index, selected_lora_info, plus edit_feedback - the string that confirms an edit went through. This node is also marked as an output node, meaning it's designed to be a terminal point in your workflow: you use it to see your library and make changes, and its job ends there. The edit_feedback output is your confirmation channel; check it after an edit rather than assuming it worked.

    Mechanically, like the rest of the family, it's reading and writing the pack's lora_tester_db.json - display pulls from the store, edits push back into it. That's also the reason to be a little careful: while the node itself is safe to run, you're now combining display and mutation in one UI, and the boundary between "I'm just looking" and "I just changed something" is one checkbox (edit_mode). Keep edit_mode off until you actually want to edit, or you may find you've been "viewing" your way into accidental changes.

    The honest verdict: this is a single-author, niche node that shows its age as a personal-project tool - the HTML-card UI is functional, not polished, and the "gallery" is really a string artifact you may need to view in a browser. But if you're living in this pack's LoRA testing ecosystem, an editable gallery beats editing JSON by hand a thousand times over. And if you're not in that ecosystem, the standalone value is low - this node's whole reason to exist is the pack's database around it.

    Install the pack and restart, then find it under loaders/lora tester:

    cd ComfyUI/custom_nodes
    git clone https://github.com/EricRollei/AAA_Metadata_System
    cd AAA_Metadata_System
    pip install -r requirements.txt
    

    Point it at your lora_list, set a min_rating so you're not staring at everything, and let search_text find what you're after. Browse with edit_mode off, edit with it on.

    Categoryloaders/lora tester

    Inputs (14)

    NameTypeDefaultDescription
    lora_listSTRINGLoRA list from LoRA Tester node
    selected_indexINT11–999Currently selected LoRA index (1-based)
    edit_modeBOOLEANfalseEnable quick edit mode for gallery cards
    gallery_sizeCOMBOmediumSize of gallery cards
    show_architectureBOOLEANtrueShow architecture badges on cards
    show_categoryBOOLEANtrueShow category badges on cards
    show_ratingsBOOLEANtrueShow rating stars on cards
    show_triggersBOOLEANtrueShow trigger words in tooltips
    filter_architectureoptCOMBOAnyFilter gallery by architecture
    filter_categoryoptCOMBOAnyFilter gallery by category
    min_ratingoptINT00–5Minimum quality rating filter (0 = show all)
    has_images_onlyoptBOOLEANfalseShow only LoRAs with associated images
    has_triggers_onlyoptBOOLEANfalseShow only LoRAs with trigger words
    search_textoptSTRINGSearch in LoRA names and trigger words

    Outputs (4)

    NameTypeDescription
    gallery_htmlSTRING
    selected_indexINT
    selected_lora_infoSTRING
    edit_feedbackSTRING