Any To IMAGE 🔽

Any To IMAGE is a type adapter from the Telegram Suite's converter family: it takes any input and declares the output an IMAGE without touching the value. Pure relabel, zero image processing. Feed it a tensor and the tensor comes out labeled IMAGE; feed it something that isn't an image and it still comes out labeled IMAGE. The node doesn't care.

Why you'd use it

The one real reason in this pack is the trigger mechanism. When you route a signal through a trigger (the suite's execution-ordering trick - see the Send Chat Action article), the trigger output is typed * (wildcard). If that signal needs to keep flowing into a port that expects IMAGE, ComfyUI's strict type checking will refuse the wildcard wire. Any To IMAGE is the bridge that re-declares the signal as an image.

That said, be honest about how often this comes up: an IMAGE is the heaviest thing you could push through a trigger, and you'd usually reach for something lighter - a seed INT, a prompt STRING, a BOOLEAN flag - as your ordering signal. If you find yourself converting a full image tensor to "any" and back just to pin execution order, it works, but you might be making your graph heavier and your diffs uglier than they need to be. The converter doesn't copy the image; the object just travels through, so the cost is in wiring noise, not memory.

The trap

Because it's a cast and not a conversion, feeding a non-image into an IMAGE port downstream can produce confusing failures - the node that consumes it will assume it's a real image tensor and do tensor things with it. If the error mentions shapes or dims, look upstream, not at this node. It declared what it was told; it didn't verify anything.

Input and output

One any input, one IMAGE output, no settings. Same shape as every converter in the pack.

Install

Shared pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/SwissCore92/comfyui-telegram-suite.git
cd comfyui-telegram-suite
pip install -r requirements.txt

Or "ComfyUI Telegram Suite" from ComfyUI Manager. No extra dependencies or downloads - the converters are bundled in the pack's converters.py. If your workflow never uses triggers, you'll likely never open this node, and that's the expected state of things.