Prompt Reader Pixaroma

What it does

Reads the metadata embedded in an image (PNG, JPG or WebP) and extracts the positive prompt that generated it. Works with ComfyUI workflows (including chained text nodes, SDXL dual encoders, and switch/combine graphs), with Automatic1111 / Forge images, and with images made on Civitai.

The readout updates the moment you pick a file, so you can read the prompt before running. If the image has no embedded prompt (a screenshot, a photo, or a file whose metadata was removed), the readout explains why.

How to use

• Click Upload Image to pick an image, or drag one onto the node.
• Or use the file dropdown to choose from images already in ComfyUI's input folder.
• The prompt appears in the text area immediately.
• Wire the text output into a CLIP Text Encode or any other text input to reuse the prompt.

Follow a connected image

Wire an image's filename into the optional filename input (for example from Load Image Pixaroma's filename output) and the node reads that image's prompt automatically. While the wire is connected it ignores its own picker, and the readout follows the connected node live as you switch images.

To go back to picking manually, just upload, drop, or pick a file on the node - that takes over and disconnects the wire.

Only images that actually have a prompt baked in (made by ComfyUI, A1111, Forge or Civitai) can be read. A screenshot or plain photo carries no prompt.

Note: the connected filename does not include the subfolder, so if you keep two images with the exact same name in different input subfolders, the run may read the wrong one. Give them distinct names to be safe (the live readout in the node always shows the correct one).

Read every picture in a folder

Wire Load Images from Folder Pixaroma's path output into filename and press Run once: the node reads the prompt saved in EVERY picture of the folder, one after the other, and sends each one on. Wire text into CLIP Text Encode and a new model makes each of your old pictures again from its own prompt.

A full path is only read from ComfyUI's input, output or temp folder, or from a folder you picked once with Browse (on Load Images from Folder, Save Image or Save Text). A folder anywhere else is refused with a message that says how to allow it.

While it is fed by a folder, the readout on the node says the prompts are read on Run; the live preview is only for a picture from the input folder.

A picture you make this way does not carry the prompt it was made from: Prompt Reader read it from another file while the workflow ran. Reading that new picture again tells you so instead of showing a wrong prompt.

Inputs

filename (optional)

A filename to read from, usually wired from Load Image Pixaroma's filename output, or a full path, from Load Images from Folder Pixaroma's path output. When connected it drives the read and overrides the picker.

Outputs

text

The recovered prompt, or an explanatory message when none was found.

The buttons on the node

Upload Image

Opens a file picker so you can choose an image and read the prompt saved inside it.

The arrows

Step to the previous or next uploaded image and read its prompt automatically.

The file name

Click it to pick a different uploaded image from a list.

Copy

Copies the prompt it found to your clipboard.