Image Info To Infotext

Image Info To Infotext takes an image_info bundle and serializes it into a single STRING in the A1111 infotext format - the Prompt: ..., Negative prompt: ..., Steps: 20, Sampler: euler, CFG scale: 7, Seed: ... text that A1111 users have been sharing inside image filenames and metadata for years. One bundle in, one human-readable text string out. It's the pack's translator between its structured metadata and the format the wider ecosystem already speaks.

Why that translation matters

ComfyUI's native workflow JSON is powerful but opaque, and it lives in PNG metadata rather than in a form people paste around. A1111 infotext, by contrast, is plain text: it travels in filenames, gets pasted into Discord, and gets read by Civitai's metadata parser. This pack saves images with A1111-compatible infotext embedded, and its Image Reader parses it back. This node is the visible middle of that loop - feed it any image_info and see exactly what the pack thinks the "text form" of your generation is.

That's genuinely useful in a few spots: previewing what will get written to metadata before you save, building a caption or notes line from a run, or just checking that a bundle restored from an old image looks the way you expect. If you've ever wanted to paste "here's the exact recipe" into a changelog, this is the node that makes the string for you.

The input and output

One optional image_info input. Leave it unconnected and you get an empty string; connect a bundle and you get the infotext. Output is infotext as a STRING, ready for a text node, a filename, or a save.

A small but important detail: this produces the public infotext. The pack supports protected/encrypted settings for sensitive values, and Image Info To Infotext emits the readable public form - which is the whole point, since the output is meant to be text people can read. If you're worried about prompt or model details leaking into a shared image, the encryption settings are the thing to configure on the save side, not something this node will paper over.

The honest limits

It's a serializer, not a round-tripper guarantee. The infotext it produces is only as accurate as the image_info it's given, and the pack's own normalization means fields it doesn't recognize may come out differently after a save/load cycle than they went in. For the core fields - prompt, negative, model, sampler, steps, CFG, seed - it's faithful, and those are the ones that matter 95% of the time.

Installing it

Ships in the ComfyUI-Info-Prompt-Toolkit pack. Via ComfyUI Manager - search "Info Prompt Toolkit" - or:

cd ComfyUI/custom_nodes
git clone https://github.com/kinorax/comfyui-info-prompt-toolkit.git
cd comfyui-info-prompt-toolkit
pip install -r requirements.txt

Restart ComfyUI. No model downloads. Small single-author pack (GPL-3.0), modest community footprint - the A1111-compatibility bridge is precisely the kind of interop the ecosystem keeps needing.