GluePuddingTool

AI Image Prompt Builder

A categorized tag library for Stable Diffusion and Midjourney. Click tags for subject, style, quality, lighting, composition, color and artist, and the tool assembles an English prompt with optional weight syntax and a negative prompt.

FreeOnline Tool
Loading…

How to Use

  1. Click the tags you want across seven categories: subject, art style, quality/rendering keywords, lighting, composition/camera, color and artist/movement (the interface labels are in Chinese, the tag values used in the prompt are English).
  2. Selected tags appear in the "Selected tags" area, where each one can be given its own weight value.
  3. Choose the weight syntax at the top: none, Stable Diffusion style (keyword:1.2), or Midjourney style keyword::1.2. The positive prompt regenerates in the chosen syntax.
  4. Click common negative tags (low quality, deformed, watermark and so on) in the "Negative prompt" area to build the negative prompt.
  5. Press "Copy positive prompt" or "Copy negative prompt" and paste the text into Stable Diffusion WebUI or Midjourney.
  6. Press "Load sample selection" to see a complete combination, or "Clear all selections" to start over.

Features

  • Seven tag categories - subject, art style, quality/rendering keywords, lighting, composition/camera, color and artist/movement - each selectable in Chinese and emitted in English.
  • Selected tags are assembled into a complete prompt in category order, with no manual stitching.
  • Individual tags can be weighted, using either Stable Diffusion's (keyword:1.2) or Midjourney's keyword::1.2 syntax.
  • A built-in library of common negative tags (low quality, deformed, watermark and more) produces the negative prompt in one click.
  • Both the positive and negative prompts can be copied with a single click.

Use Cases

Generating images in Stable Diffusion WebUI
Paste the generated text straight into the positive and negative prompt boxes in WebUI to assemble a strong prompt quickly.
Tuning Midjourney prompts
Use Midjourney weight syntax (::1.2) to fine-tune how strongly each tag influences the image and iterate faster.
Learning AI art prompting
The categorized tag library shows newcomers which elements make up a complete prompt - subject, style, quality, lighting, composition and color.
Building a team prompt library
Copy and store tag combinations that reliably produce good results as reusable starting templates for your team.

FAQ

Why are all the tags in English?
Mainstream text-to-image models such as Stable Diffusion and Midjourney generally understand English prompts better than Chinese, so the tag values are all English while the interface offers Chinese labels to help you choose.
Do I have to use weight syntax?
No. With the default "no weights" setting, every tag is simply joined at equal weight. A weight marker is only added when you change a tag's weight away from 1 in the "Selected tags" area and have picked the SD or MJ syntax.
Which weight syntaxes are supported?
Two: the Stable Diffusion style (keyword:1.2) and the Midjourney style keyword::1.2, switchable from the dropdown at the top.
Are my generated prompts saved?
No. Nothing is saved automatically - your selections only exist in the current page session, so use the copy buttons to keep results in your own notes.