📥 Load Image Url

The stock Load Image node makes you upload a file first. This one skips that: paste a URL and it downloads the image and hands it straight into your workflow as an IMAGE. It's a small quality-of-life node, but it removes a genuinely annoying step - no more right-click-save-then-upload just to run a reference picture through img2img or a ControlNet.

Where it earns its keep

The obvious win is speed. You found a reference on the web, you want it in ComfyUI, and you don't care about keeping the file - paste the link, done. But the bigger payoff is automation. Because the source is a text field, you can drive it from other nodes: pipe in a URL from a text file, a spreadsheet, or a prompt, and batch through a list of image links without touching the file browser once. That turns "load this one picture" into "process these hundred URLs," which is the kind of thing you can't do with an upload-only loader. If you're building an automated pipeline, a URL-driven image input is a real unlock.

The inputs and outputs that matter

• Url - the one input. A direct link to an image. Paste it, run, and the node fetches and decodes it.

The single output is IMAGE, which wires into anything expecting an image: a VAE Encode for img2img, a preprocessor for ControlNet, an upscaler, whatever you'd normally feed a loaded image into.

Installing it

ComfyUI Manager: search SDVN_Comfy_node, install, restart. Or by hand:

cd ComfyUI/custom_nodes
git clone https://github.com/StableDiffusionVN/SDVN_Comfy_node

Then pip install -r custom_nodes/SDVN_Comfy_node/requirements.txt from your ComfyUI root, and restart.

Where people get burned

The URL has to point at the actual image. A link to a webpage that displays an image is not a link to the image file - you want the address that ends in the image itself (right-click, "copy image address"), not the page URL. Feed it a page and the fetch either fails or downloads HTML.

Hotlink protection and logins. Some sites block direct fetches or require a session cookie, so a link that opens fine in your browser can still 403 the node. If a URL won't load, that's usually why - download it manually and use the normal loader instead.

Dead links break the run. Because the image is fetched at runtime, a URL that goes offline (or rate-limits you) fails the whole workflow, where a local file wouldn't. For anything you'll rerun a lot, or a picture you care about keeping, download it once and load it locally - the URL loader is best for throwaway references and automated batches, not for assets you depend on.