2🐕Internet celebrity filter

The name translates roughly to "internet celebrity filter," and once you look at the choice list - 1977, Aden, Brannan, Brooklyn, Clarendon, Earlybird, Gingham, Hudson, Inkwell, Kelvin, Lark, Lofi and more - it's obviously a set of the classic Instagram/CSSgram-style photo presets. If you want that specific "shot on an old iPhone filter app" look on a generation without touching a color-grading node yourself, this is the one-click version.

How it works

You feed it an image and pick a style from the 26-filter list; it applies that preset's color curve and outputs the filtered result. Each name is a recognizable preset from the Instagram filter era - Clarendon boosts contrast and cools shadows, Gingham flattens and warms everything, 1977 pushes a warm, slightly overexposed vintage tone - so if you've used Instagram or a CSSgram-based filter tool before, the names will already mean something to you. There's also an All toggle (default off); with it on the node runs every filter in the list rather than just the one you picked, which is the batch-testing move - generate one image, run it through all 26, and eyeball which preset actually works for that shot before committing.

The inputs and outputs that matter

• image (IMAGE) - required, the source frame.
• style (enum, 26 choices) - which preset to apply. This is the only thing most people will touch.
• All (BOOLEAN, default false, optional) - flip it on to run every filter instead of just style, useful for comparing presets side by side.

Output is a single IMAGE - wire it straight into a Save Image or preview node.

How to install it

ComfyUI Manager: search Comfyui-ergouzi-Nodes. Manual install:

cd ComfyUI/custom_nodes
git clone https://github.com/11dogzi/Comfyui-ergouzi-Nodes.git

Restart ComfyUI. Heads up: the author's README explicitly says this English-titled repo is frozen - no further updates - with active development continuing on a Chinese-named sibling pack. What ships on comfy.icu is the frozen snapshot.

Common issues & troubleshooting

Filters look identical to each other, or don't look like Instagram's. These presets are re-implementations, not licensed from Meta - expect the general character of each look (Clarendon = punchy contrast, Gingham = washed pastel) rather than a pixel-perfect match to the app version. If two filters look too similar for your taste, that's a reasonable outcome given how many of the 26 share the same underlying "warm vintage" or "cool flat" family.

Running All on a large image is slow and produces 26 outputs. That's the trade for the convenience - treat it as a preview/comparison step on a downsized test image rather than something you leave on for your final batch render.

The look fights your model's own color tendencies. Filters like this operate purely in pixel space after generation - they don't know or care what checkpoint made the image. If your base model already leans warm or desaturated, some presets will double down uncomfortably. Generate first, then filter, and judge each preset against the actual image rather than assuming a preset that looked good on one shot will generalize to the next.