🔱 Artha Image Save

Artha Image Save is the 🔱 Artha pack's disk-writing node: it takes an image and writes it to ComfyUI's output folder as a PNG. Its one distinguishing feature is a compression slider (0–9) that the core save node doesn't expose - you get real control over file size versus save time.

Let's be honest about the compression knob first, because people overestimate it. PNG compression is lossless, so dropping from 9 to 4 doesn't hurt image quality at all - it only trades CPU time for file size. ComfyUI's default save is already good; the slider is a "shave a few MB off a big render" convenience, not a quality control.

How it works

The node takes an IMAGE input, builds the save path from a filename_prefix, and writes the PNG into ComfyUI/output/ using ComfyUI's standard save helper with the chosen compression level. Standard ComfyUI save behavior applies: same prefix, files get auto-numbered to avoid clobbering, and the workflow metadata gets embedded in the PNG - which is how "workflow included" sharing works in this ecosystem.

Inputs that matter

• images - the IMAGE input to save.
• filename_prefix - optional string, defaults to ComfyUI. Change it to something project-specific so your outputs don't all pile into one folder naming scheme.
• compression - integer 0–9, default 4. Lower = faster, bigger file; higher = slower, smaller file. For a 1024×1024 render the differences are usually in the single-digit MB range.

It's an output node, so no outputs - it ends the graph path and dumps to disk.

Installing it

Same one-pack install as all Artha nodes:

cd ComfyUI/custom_nodes
git clone https://github.com/Cyrostar/ComfyUI-Artha-Nodes
pip install -r ComfyUI/custom_nodes/ComfyUI-Artha-Nodes/requirements.txt

then restart ComfyUI, or install "ComfyUI-Artha-Nodes" through ComfyUI Manager. No API key, no model files - this node is pure Pillow/ComfyUI plumbing.

Common issues

• Can't find the file: it lands in ComfyUI/output/ (or wherever you've remapped output), named {prefix}_{number}.png, not in the node itself. Check the prefix you set.
• Slow saves on huge upscales: you're running high compression on a multi-megapixel PNG. Drop it to 1–4 - you'll lose nothing visually since PNG compression is lossless.
• Not saving anything: no image wired to images, or the node sits on an unexecuted branch.

Where it fits: if you're building an Artha-only workflow chain, this is the terminal node that gets your result onto disk, and the compression slider is a small but real quality-of-life win over the core save node. If you already have the core SaveImage node and never think about PNG size, you're not missing much by skipping it.