[Free Tool] I built a lightweight, local HTML Prompt Randomizer

Hey everyone,
Yes i know that wildcards exists, but somehow that got too complex for me within the prompt (Following article and tool made with Support iof Google Gemini)
I’ve been doing a lot of local AI generation in ComfyUI and I needed a quick, lightweight way to generate randomized prompt batches and test different variables. I vibecoded a Prompt Randomizer that runs entirely offline as a single .html file.
- Weighted Randomness (Roulette Wheel) You can define variables like Haircolor and add weights to them, similar to standard prompting syntax. Example: Blue/Blonde:3/Brown:0.5. The tool will automatically roll "Blonde" much more often than the others.
- Consistency Locks (The <*Tag> Feature) If you have a prompt where a character is mentioned multiple times, you don't want their hair color changing mid-prompt. If you use <*Haircolor> instead of <Haircolor>, the tool rolls the color exactly once and locks that exact value for all other <*Haircolor> and <Haircolor> tags in the same prompt.
- Variation Batching (Great for Grids / Testing) Want to test 5 different lighting setups or 10 camera angles without changing the rest of the prompt? Select a variable (e.g., Lights), and hit Variation Batch. The tool will roll a fixed base prompt and iterate through every single value of your selected variable. It exports a .txt with the prompts, and a second .txt with just the Labels. Attached is a workflow why, The second Labels can be used as an Overlay text to know what should change.
- QoL Features
Right-Click Context Menu: Just right-click in the text area to quickly insert your defined tags.
Batch Generation: Generate 5, 50, or 1000 randomized prompts with one click and download them as a
.txtfile. I attached a workflow which Loads One Line of the Txt into the final string.Save/Load Configs: Save your complex variable setups as a
.cfg(JSON) and load them whenever you need them. Feel free to share yours ^^
Example Template:
<Cameraangle> <Lights> photo of a <Age> old man with <*Haircolor> Hair standing next to his <Age> year old wife, also with <*Haircolor> Hair in <Location>.
Output:
Close-Up golden hour sunlight photo of a 30 old man with Blonde Hair standing next to his 25 year old wife, also with Blonde Hair in Supermarket.
How to use it:
I have attached the relevant Workflows and Files, CivitAI does not allow .html files, so its a .txt but if you rename the extension to HTML it should work :)
Otherwise Here are the Pastebin-Links:
The HTML itself: https://pastebin.com/Pp5a9de0 (click download and open via your browser)
Workflow for Random-Prompt: https://pastebin.com/JFVPgKaG ( click download, drag and drop into ComfyUI, change the FilePath in the QUEUEJOV-Node . I used the Jovimetix-Node because otherwise the .txt does not refresh once you recreate the batch from the HTML.
Workflow for Variation-Batch: https://pastebin.com/dbWnUE0s (click download, drag and drop into ComfyUI, change the FilePaths of the Variationbatch.txt and the Variationslabels.txt to your Filepath. The Variationslabels.txt is only needed for the Text-Overlay, you may not need it.
Example-Config https://pastebin.com/8bftpfvH (click download, in the HTML click load Config, its okay to be a txt-data)
Let me know what you think or if you have any ideas for new features!