TT Image Preview / Save

Every ComfyUI graph ends somewhere, and TT Image Preview / Save is a fine place to end it. Most people run a native PreviewImage and a SaveImage side by side, which means your output shows up twice and your disk fills with two copies of everything. This node does both jobs from one box: it previews the image in the graph and writes it to disk, with a save_image toggle if you want to go preview-only for a while.

The details that make it worth using over the native pair are in the filename handling. filename_prefix is your base name (default tenser-tensor), and filename_format picks how the file gets numbered:

• name-### - auto-increments like tenser-tensor-1.png, tenser-tensor-2.png. The counter scans the output folder, so it never overwrites.
• date-name-### - prepends the date, useful when you save across multiple sessions.
• name-datetime - a timestamp to the second; every render gets a unique file, which is what you want for batch runs you might want to diff later.

subfolder lets you tuck results into output/<subfolder>/ - handy when you're saving different project streams into the same workflow. And there's a real safety feature hiding here: the code refuses to save outside the output directory. Try to escape with ../ and it throws InvalidSavePathError instead of scattering files across your drive.

image_format and the quality knobs behave the way you'd expect from any image tool, with one quirk: PNG uses compression_level (0–9, default 9 for smallest files) and ignores image_quality, while JPEG and WEBP use image_quality and ignore compression. So if you're exporting JPEGs for a client, the number that matters is image_quality; for PNG archives it's compression_level. All three formats come from the pack's own FORMAT_EXT map and are handled with PIL under the hood - nothing exotic, which is why this node is reliable on any install.

Install is the pack standard:

cd ComfyUI/custom_nodes
git clone https://github.com/tenser-tensor/ComfyUI-TenserTensor

or search "TenserTensor" in ComfyUI Manager and restart.

Where people get confused:

• "Where did my image go?" The pack saves to ComfyUI's output/ directory (with your subfolder appended), not somewhere exotic. If you can't find it, check ComfyUI/output/.
• "Why is my PNG 40 MB?" Because compression_level default is 9, that should make it smaller - but if you've set it to 0 you're getting the largest, fastest-write PNG possible. Set 9 for archival, and if you see no filename in the preview widget, you've probably left save_image off and it's only previewing.
• No outputs, and that's fine. This is an output node - it returns nothing to wire onward; it's the end of the line. That's expected.

One caveat that applies to the whole pack: TT_ImagePreviewSave is the original V1 node and is marked deprecated while the author migrates to API V3 (the TT_ImagePreviewSaveNode replacement). It works fine now, but if you're building something you'll maintain for a year, the V3 sibling is the safer bet.