Coordinate_loadImage

Normally, marking points on an image in ComfyUI means two nodes: LoadImage, then something to pick coordinates on top of it. Coordinate_loadImage collapses that into one step - pick your file from the input folder and click your points in the same node.

What it is and why you'd reach for it

This belongs to ComfyUI-Apt_Preset's ImgCoordinate family, the cluster of nodes in cardenluo's utility pack built around passing image regions and points as portable JSON data. If you're building a workflow where the first thing that happens to an image is "mark some points on it" - annotating reference photos, tagging face or object locations before a downstream crop or conditioning step, prepping test images for an automation pipeline - this saves you a node and a wire versus chaining core LoadImage into Coordinate_fromImage separately.

How it works

The image field is an enum picker pulling from ComfyUI's input directory, exactly like core LoadImage - pick a file, it loads. points_data works the same as it does across the rest of this node family: a JSON string of {x, y, index} points, defaulting to an empty list here ([]), that the node's own preview lets you build up interactively by clicking on the loaded image.

Inputs and outputs that matter

• image (required, enum) - which file to load, picked from ComfyUI's input folder.
• points_data (required, STRING) - the point list, starting empty; click on the preview to populate it, or hand-edit the JSON directly if you already know your coordinates.

Outputs: ORIGINAL_IMAGE - the loaded image as-is, and img_tensor - a second image output, useful when you need the decoded tensor to flow one direction (into your actual generation pipeline) while keeping the untouched original available separately (for reference or display). Plus coordinates, the JSON point list, ready for Coordinate_MarkRender, Coordinate_pointCombine, or Coordinate_SplitIndex downstream.

Installing it

Through ComfyUI Manager: search ComfyUI-Apt_Preset, install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/cardenluo/ComfyUI-Apt_Preset.git

Run install.bat on Windows for dependencies; on Linux/Mac, install the pack's Python requirements yourself since no equivalent script is bundled. No model downloads needed - this is file I/O and a click-based UI, not inference.

Common issues

If your image loads fine but clicking does nothing, that's consistent with the node's frontend JavaScript not loading - which happens when the pack's backend import failed on startup (see below) and the accompanying web assets never registered. Check the browser console as well as the ComfyUI server log if the click surface seems dead.

Because points_data starts as an empty array here (unlike Coordinate_fromImage, which defaults to one centered point), downstream nodes expecting at least one coordinate will get nothing if you forget to click before running the graph - worth a sanity check before you queue a batch.

And, the standing caveat for this whole pack: ComfyUI-Apt_Preset is large and actively developed, and at least one user has reported an IMPORT FAILED error at ComfyUI startup after installing it. If Coordinate_loadImage - or any Apt_Preset node - isn't appearing in your node search, check the startup console log for the actual missing-dependency error before reinstalling.