🪠️ WWAA Save JPEG

ComfyUI's default image save writes lossless PNGs, and that's the right call for keeping your work pristine. But PNGs are giant, and the moment you're producing web images, avatars, or anything that'll be uploaded somewhere, you want JPEGs at a sane quality - which stock ComfyUI makes you reach for an extra pack to do. WWAA Save JPEG is that extra pack's answer: save your output directly as JPEG with a quality slider, auto-incrementing filenames, and date-stamped subfolders.

How it works

Feed it images (a batch is fine - each image gets its own numbered file), set filename_prefix and quality, and it writes prefix_00001.jpg, prefix_00002.jpg, … into ComfyUI's output directory. Two details make it more useful than a plain save:

• Numbering is collision-aware. It scans the target directory, finds the highest existing number for your prefix, and continues from there - so re-running doesn't overwrite your earlier files. (New prefix, new series starting at 00001.)
• %date:format% and subfolders. The same date-token syntax as the pack's Metadata Saver works here: %date:yyyy-MM-dd%/myfile creates a dated folder and saves inside it. Prefix with a slash, and the folder is created for you.

The quality input (1–100, default 95) is the real reason to pick this over converting later. JPEG compression is a visible tradeoff - default 95 keeps things basically lossless-looking for a large-ish file; dial to 70–80 for web and watch the size drop. Pair this node with the pack's JPEG Preview (which tells you the actual KB at a given quality) and you can tune before committing.

Install

It's in the WWAA-CustomNodes pack. ComfyUI Manager: search "WWAA Custom Nodes", install, restart. Or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/hgabha/WWAA-CustomNodes
# restart ComfyUI

No extra dependencies - the pack needs only ComfyUI's built-in numpy/Pillow/torch, and this node is pure Pillow under the hood. It lives under the 🪠️ WWAA/image menu.

Notes and gotchas

The conversion is a straight Pillow RGB save, so you get a standard baseline JPEG - no chroma subsampling options, no progressive mode, no EXIF. If your workflow needs photographic-quality JPEG control, a dedicated encoder pack beats this; if you just want "save a decent JPEG, numbered, organized," it's perfect. One caveat: numbering depends on scanning for files that start with your prefix, so changing the prefix mid-run starts a new series - use the date token in the prefix and you'll always have clean, chronological folders instead of one ever-growing pile. It's a small, opinionated tool from the WeirdWonderfulAI.Art team, built because their batch workflows needed JPEGs that didn't explode in size - and it does exactly that.