rockbenben/img-prompt

What it solves

IMGPrompt is a visual prompt editor designed for users who do not think in English or struggle to find the specific technical terms (like "Tyndall effect" or "Fisheye Lens") that AI image and video models require for high-quality results. It bridges the gap between a user's visual intent and the English keywords needed by models like Midjourney, FLUX, and Veo.

How it works

The tool provides a curated library of over 5,000 bilingual tags across 16 subjects and 212 categories (such as lighting, composition, and art styles). Users select a subject and category, then click on tags—many of which include preview images—to assemble an English prompt in a real-time panel. It also includes a translation box that converts free text from various languages into English, suggesting curated tags based on the input.

Who it’s for

Artists, designers, and AI art enthusiasts who want precise control over their generated images and videos without needing to be fluent in English or having to memorize technical photography and art terminology.

Highlights

• Multilingual Support: Interface and tag library translated into 18 languages.
• Visual Tagging: Nearly every tag has a hoverable preview image to show the effect of the style. -- Broad Model Compatibility: Works with GPT-Image-2, Midjourney, FLUX, Veo, Kling, Stable Diffusion, and DALL·E.
• Privacy-Focused: No backend; user selections are stored in the URL and browser.
• Extensible Data: The tag library is stored as plain JSON files, making it easy for contributors to add or edit terms.