Lmcq Image Saver Transit

This one's a single-letter difference from LmcqImageSaver that matters a lot in practice: same watermarking, format, and enhancement pipeline, but it hands the image back out afterward instead of ending the branch. If you've ever wanted to save a watermarked file and keep using that image downstream - feed it to a preview node, upload it, run it through another pass - this is the node, and the plain LmcqImageSaver can't do it because it has no outputs at all.

Why this needs to be a separate node. ComfyUI treats output nodes (the ones that write files) and pass-through nodes differently, and normally you can't have both in one - SaveImage ends its branch, full stop. LMCQ's answer is to ship a second copy of the image-saver logic that also happens to expose an images output, so you can literally treat it as a mid-workflow checkpoint: save-and-continue instead of save-and-stop. It's not a smarter node than LmcqImageSaver, just a differently-shaped one.

The fields - identical to LmcqImageSaver. images in, filename_prefix (default LMCQ) for naming, format (png/jpg/webp) and quality (1-100, only relevant for jpg) for the file itself. Watermarking is off by default via apply_watermark; flip it on and set watermark_type (text or image), watermark_text, watermark_size, watermark_position (one of nine anchor points), and watermark_opacity. Wire an image into the optional watermark_image input if you're stamping a logo rather than text. enhance_brightness and enhance_contrast do a light post-process pass (both default to 1, meaning no change), and save_metadata controls whether the ComfyUI workflow gets embedded in the saved file.

The difference that actually matters is the output: images, type IMAGE, wired straight back out so you can chain more nodes after the save.

When you'd actually use this over the plain saver. A common pattern: save a watermarked, compressed JPEG as your "public" deliverable, then continue the same branch into a PreviewImage node so you can eyeball the watermark placement in the ComfyUI UI without opening the file on disk. Or save an intermediate step in a multi-stage workflow (upscale pass, inpaint pass) while still needing the image for the next node. The plain saver forces you to either duplicate the image with a separate node or accept that the branch ends there - this one just lets you keep going.

Installing it. Same as the rest of the pack - through ComfyUI Manager (search "ComfyUI-LMCQ") or:

cd ComfyUI/custom_nodes
git clone https://github.com/sebord/ComfyUI-LMCQ

Restart ComfyUI to pick it up. No extra dependencies for this node specifically.

Things to watch for. Same PNG/quality trap as its sibling: the quality slider does nothing unless format is set to jpg, so don't be surprised if a PNG export ignores it entirely. And because this node is doing real image processing (watermark compositing, brightness/contrast) before passing the tensor back out, downstream nodes see the processed image, not the original - if you need both the raw and the watermarked version, split the branch before this node rather than after it.

The pack-wide landmine applies here too: ComfyUI-LMCQ's encryption nodes depend on a compiled runtime module that's occasionally missing on fresh installs, throwing a ModuleNotFoundError that crashes the whole pack's import - meaning this node (and everything else in LMCQ) simply won't appear in your node list. If that happens, it's a known, reported issue; the README's January 2026 changelog entry has the fix (copy in the Python-version-matching runtime files, drop the rest, restart ComfyUI).