Load Image Clipboard
Grab a reference image straight from your clipboard, no file round-trip
- IMAGE
- MASK
You found the perfect reference in Discord or on a design site, and your instinct is to copy it and paste it into ComfyUI. The stock Load Image node can't do that - you save the file, hunt it down in your downloads, then find it in the dropdown. Load Image Clipboard is stock Load Image with a "Use clipboard image" button bolted on, so the copy-paste instinct just works. The name sounds like it might be calling some external service; it isn't. Everything happens locally against your own ComfyUI.
How it works
The node itself subclasses ComfyUI's LoadImage, so the core is stock. The interesting part is the button, which is front-end JavaScript. When you click it, the browser reads an image off your clipboard (navigator.clipboard.read), uploads it to ComfyUI's input directory through the normal /upload/image API under a timestamped name like clipboard_20260816_143022.png, then sets the node's image dropdown to that file and runs. The extension picks the file extension from the mime type, so a JPEG pasted from a site stays a .jpg instead of being re-wrapped as a PNG.
If the browser refuses direct clipboard access - which happens, more on that below - it falls back: an alert tells you to press Ctrl+V, and it listens for a paste event for up to 30 seconds, then uploads that.
The inputs that matter
- image - not a value you set, but the dropdown of files in your
ComfyUI/inputfolder. After you hit the button it'll show the freshly uploaded clipboard file, so you can re-run with it without re-pasting.
Outputs are IMAGE and MASK, exactly what stock Load Image gives you. Wire IMAGE into your VAE encode or img2img chain, and MASK into whatever masking step you need - same behavior you already know from the stock node.
Why you'd reach for it
Any time the image you want is already on your clipboard and you'd rather not touch a file manager: a reference grabbed from a website, a screenshot you just took, an image pasted into Discord that you want to img2img. It also plays nicely as an input-staging node - after it runs, the file exists in input for reuse.
How to install it
ComfyUI Manager β search "ComfyUI AstroCorp Nodes" β install β restart. Or:
cd ComfyUI/custom_nodes
git clone https://github.com/AstroCorp/ComfyUI-AstroCorp-Nodes
Restart, and it's under "ππ AstroCorp ππ". No models to download; the pack's pyproject.toml lists zero Python dependencies (the tensorrt line in requirements.txt is for the pack's ONNX upscaler nodes only - you don't need it here).
Common issues
The clipboard button's number-one failure is browser policy. Reading the clipboard directly is only allowed in a secure context - https or localhost - and only when the page is focused. If ComfyUI is running on plain http at a LAN IP, or the tab isn't active, the button falls back to "press Ctrl+V now." That fallback exists for a reason: use it. It's not broken, it's the browser being cautious.
One thing to know: every click uploads a new clipboard_*.png into your input folder, and they accumulate. It's harmless, but if you're fastidious you'll want to sweep them out occasionally.
For a tiny personal pack with no community footprint yet, the mechanism is actually well thought out - timestamped names, mime-aware extensions, graceful fallback. It does one thing and does it locally. If you live in the copy-paste workflow, it's worth the two-minute install.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| image | COMBO | 1 options: example.png |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| IMAGE | IMAGE | β |
| MASK | MASK | β |