The prompt resources for IMGPrompt are collected from online sources, including promptoMANIA and sd-webui-prompt-all-in-one. After deduplication and standardization, nearly 5,000 prompts have been compiled. These can be used in various AI image generation and text-to-video scenarios.
All prompt data is stored in the project directory under src/app/data with the following structure:
Each prompt corresponds to a JSON object with the following structure:
Explanation:
| Field | Required | Description |
|---|---|---|
displayName |
✅ Yes | Unique English prompt |
langName |
✅ Yes | Name displayed in the UI (localized or custom translation) |
object |
✅ Yes | Main category or type, e.g., "Character", "Style", etc. |
attribute |
✅ Yes | Subcategory or descriptive tag for further classification |
description |
❌ No | Tooltip text shown on hover to provide more context |
displayName unique; duplicates are ignored and make suggestions noisy.object/attribute values across languages to keep navigation aligned.langName concise so tag chips stay readable.To expand the prompt list, you can use the IMGPrompt Data Converter to convert your custom data into the standard format. The converted results should be saved to the path src/app/data/prompt-custom.json. This file is automatically loaded at app startup and applies to all language settings.
Here is an example of the prompt-custom.json format: