📂 Load Image (Multi 10)
Load Image (Multi 10) for big reference sets
- image_1
- image_2
- image_3
- image_4
- image_5
- image_6
- image_7
- image_8
- image_9
- image_10
Load Image (Multi 10) is the bigger sibling of the pack's Load Image (Multi 5) node, and honestly that's the whole story. Same author, same pack, same install, same design - just twice the slots. In the source it's literally a subclass with OUTPUT_COUNT = 10 instead of 5. Everything the 5-node does, this does too: each of the ten slots has a file dropdown (image_1–image_10, listing your input folder with upload support) plus a URL box (url_1–url_10), the URL wins if you fill both, an empty slot outputs a 64×64 black tile by design, and on_error defaults to failing loudly with the reason shown on the node rather than spitting out a silent black image. Ten IMAGE outputs, image_1–image_10, each a separate [1, H, W, 3] tensor.
When do you actually need ten?
Rarely, is the honest answer - but when you do, nothing else is this convenient. The genuinely wide cases: a multi-unit ControlNet stack with several condition images, IPAdapter setups pulling multiple style or identity references, character-consistency graphs that keep the same faces across several adapters, or batch img2img where you're manually feeding a folder's worth of source images. If you only need four references, the 5-slot node is the cleaner pick - fewer sockets to trip over, a smaller node on the canvas. Get the 10 when the workflow is genuinely wide, or when you're sharing a template and don't want to rebuild it when someone needs a ninth input.
The honest downside is width. Ten slots means a wall of sockets and ten wires leaving the node; on a busy canvas it gets hard to read, and legibility is exactly what separates a usable graph from spaghetti (the KB's plumbing notes on routing-vs-legibility apply here). That's the trade-off for not stacking five Load Image nodes vertically.
Install and gotchas
cd ComfyUI/custom_nodes
git clone https://github.com/heyzne/comfyui-image_psss.git
Restart ComfyUI, or search "comfyui-image_psss" in ComfyUI Manager. Dependencies are torch, numpy, and Pillow - already present in every ComfyUI. No models to download.
Same failure modes as the rest of the pack, worth repeating because they bite: dead or expired URLs should now turn the node red with a reason - if you're on an old version and a URL silently yields black, git pull to get the v2.1 fix; 403 hotlink protection can still block a fetch (the pack sends a browser User-Agent but no referer), in which case save the file to input and use the dropdown; and don't confuse the black 64×64 tile (empty slot, normal) with the magenta one (that slot failed in placeholder mode). Because both Multi nodes share everything except slot count, picking wrong costs you nothing but swapping the node - so err on the side of the 5-slot unless your reference set is really that big.
Inputs (21)
| Name | Type | Default | Description |
|---|---|---|---|
| image_1 | COMBO | 2 options: (none), example.png | |
| url_1 | STRING | — | |
| image_2 | COMBO | 2 options: (none), example.png | |
| url_2 | STRING | — | |
| image_3 | COMBO | 2 options: (none), example.png | |
| url_3 | STRING | — | |
| image_4 | COMBO | 2 options: (none), example.png | |
| url_4 | STRING | — | |
| image_5 | COMBO | 2 options: (none), example.png | |
| url_5 | STRING | — | |
| image_6 | COMBO | 2 options: (none), example.png | |
| url_6 | STRING | — | |
| image_7 | COMBO | 2 options: (none), example.png | |
| url_7 | STRING | — | |
| image_8 | COMBO | 2 options: (none), example.png | |
| url_8 | STRING | — | |
| image_9 | COMBO | 2 options: (none), example.png | |
| url_9 | STRING | — | |
| image_10 | COMBO | 2 options: (none), example.png | |
| url_10 | STRING | — | |
| on_erroropt | COMBO | 报错提示 (推荐) | 加载失败时:报错提示(节点变红显示原因)或输出品红占位图 |
Outputs (10)
| Name | Type | Description |
|---|---|---|
| image_1 | IMAGE | — |
| image_2 | IMAGE | — |
| image_3 | IMAGE | — |
| image_4 | IMAGE | — |
| image_5 | IMAGE | — |
| image_6 | IMAGE | — |
| image_7 | IMAGE | — |
| image_8 | IMAGE | — |
| image_9 | IMAGE | — |
| image_10 | IMAGE | — |