Bria Multi Image Select

Here's the plot twist: despite living in a pack called BRIA AI API nodes, this one calls no API, needs no key, and costs nothing. Bria Multi Image Select is a plain local file picker that reads several images out of your ComfyUI input folder and hands them to the graph as one batch. It exists because BRIA's API nodes are built to process batches - feed them a stack of images and they'll run each one through the API in a loop - but ComfyUI's built-in Load Image only grabs a single file at a time.

So the workflow you're actually building is: Multi Image Select → any Bria API node (Replace Background, Enhance, Eraser, the FIBO nodes) → Save. It's the front door for "process this folder of product shots" workflows, which is a very BRIA thing to want to do, given their target audience is eCommerce teams.

How it works

The node shows a Select Images button (backed by a small JS widget in the pack's web/ folder). Click it, pick however many files you want from the input directory, and the button writes the selected paths into the selected_paths string field as JSON. On execution the node parses that list, opens each image with PIL, and returns them.

The two outputs

• images - an IMAGE output holding all the selected files as a single batch, original sizes preserved.
• filenames - a comma-joined STRING of the base filenames. Handy if you want to track which API result came from which source image, or feed a naming convention downstream.

That filenames output is easy to overlook and genuinely useful: the API nodes return results in the same order you fed them, so the string tells you what slot 3 actually was.

Install

Same as the rest of the pack - it's already there if you have it:

cd path_to_comfyui/custom_nodes
git clone https://github.com/Bria-AI/ComfyUI-BRIA-API.git

or search "BRIA API" in ComfyUI Manager, then restart. No extra dependencies, no models, no token.

Gotchas

The picker only looks in the standard input directory, so drop your files there first. If you run the node with nothing selected, it raises "No images selected" - the button is the input, the string field is just where it stores the result, so don't try typing paths into it manually (though multiline JSON pasted in does work, if you're into that). And since there's no API involved, it's the one node in this pack that's completely free - don't be surprised when your ComfyUI queue blasts through it instantly.

One more thing worth knowing: the images keep their original size, so if your batch is mixed-resolution you may want an upscale/resize node between this and the API call, since some of BRIA's endpoints behave better on uniform inputs.