New ComfyUI Node I created that loads images and automatically extracts prompts from multiple sources. Handy for working with AI-generated images from Civitai, datasets, and local files where you want to remix/reuse the prompt from the image, or the seed/steps that are stored within the metadata.

Use Cases:

1. Reproducing images from civitai or other sources that contain metadata

2. Iterating or slightly modifying images with the same prompt and settings but you can add string merging nodes to combine your prompt with the original prompt

This webUI is also compatible with Gallery-DL images that contain metadata, and will look for caption files in the same folder the gallery-dl civitai image is.

See a more advanced workflow I made that incorporates this new plugin:

https://civarchive.com/models/1671988/nunchaku-flux-img2img-depth-outpaint-hr-fix-upscale-detailer-civitai-image-and-prompt-loader

https://github.com/mikf/gallery-dl

• Civitai Integration: Load images directly from Civitai URLs with automatic metadata extraction (if the image metadata exists in the API response for that image)

• 📁 Local File Support: Load from local paths or ComfyUI's input folder

• 🔍 Smart Prompt Detection: Automatically extracts prompts from:

  • Civitai API metadata (primary source for Civitai images)

  • Caption files (.txt files with same name as image)

  • EXIF metadata (from generation parameters within the image itself)

• ⚙️ Complete Generation Info: Extracts seed, steps, CFG scale, and both pos/neg prompt

See Github for installation, or use ComfyUI Manager to install.

https://github.com/hassan-sd/comfyui-image-prompt-loader

Basic Setup

1. Add the "Image & Prompt Loader" node to your workflow

2. Choose your input mode:

   • Local Upload: Use ComfyUI's standard image upload

   • Local Path: Specify a full file path

   • Civitai URL: Load directly from Civitai

Civitai Usage

1. Set input mode to "Civitai URL"

2. Paste a Civitai image URL (e.g., https://civarchive.com/images/123456)

3. Optionally add your Civitai API token for private images

4. Run the node - it will:

   • Fetch metadata from Civitai API

   • Download and display the image

   • Extract prompt, negative prompt, seed, steps, CFG scale

Local File Usage

1. Set input mode to "Local Path"

2. Enter the full path to your image file

3. The node will look for caption files in the same directory

4. Supports common dataset structures and naming conventions