<h2 id="about-this">About This</h2>This is a custom node I've made which adds a simple utility for running a whole folder of text prompts automatically. It should work with any model or text to image workflow, so you can use it with Krea2, Flux, SDXL, Illustrious, Pony, or pretty much anything else.The main part is the custom ComfyUI-FolderPromptLoader node. Instead of manually copying and pasting prompts into your workflow one by one, you simply point the node to a folder containing <code>.txt</code> files. It loads each prompt and sends them through the workflow automatically, creating a queue entry for every text file.It also sorts the files in natural order, so <code>prompt2.txt</code> runs before <code>prompt10.txt</code>, which makes numbered prompt collections much easier to manage.<h2 id="installation-and-guide">Installation &amp; Guide</h2>Unzip <code>ComfyUI-FolderPromptLoader</code> into <code>ComfyUI/custom_nodes/</code>, restart ComfyUI and refresh the browser.<ol><li>Open the workflow</li><li>Set <code>folder</code> to your prompts folder. Absolute or relative paths both work, and pasted paths with quotes are fine.</li><li>Click Queue all files. You get one queue entry per text file, in natural order, so <code>prompt2</code> runs before <code>prompt10</code>.</li></ol>If the button doesn't show up or work, leave <code>control_after_generate</code> on <code>increment</code>, set the Run count to the number of files, and press Queue.This is mainly designed as a small quality of life utility for anyone working with large prompt collections and wanting to batch run them without manually loading every prompt.<h2 id="where-to-get-prompts-(tip)">Where to get prompts (tip)</h2>Well, of course, you can copy and paste each prompt from your favorite LLM or write them manually, but there is an easy way to mass produce prompts by using CivitAI itself.Simply go to the Training a LoRA section of the website, choose any options and any names, then hit Next. In the second step, where you have to source your images, upload as many images as you want and use Auto Label.It's pretty much free and uncensored. Once it's finished, you don't need to proceed with the actual training. Just use the Download as ZIP file option. It will download both the captions and the images.Unpack the downloaded file and you'll have a folder full of prompts, ready to be loaded directly into this node.

Files with SHA256: f8e38ee54675c11d0b417a02d5af6cc8f577182f6d72cb168e66e9f749514aec