IMGPrompt · Free Visual Prompt Builder

IMGPrompt is a prompt editor for text-to-image models that helps you compose English prompts through visual tags with bilingual labels. It officially supports Nano Banana, GPT-image-2, Midjourney, FLUX, Seedance, Veo, and Kling for image and video; the generated prompts also work with Stable Diffusion, DALL·E, and any model that accepts compatible syntax.

Completely free, no registration or login required. All prompt composition and input happens locally in your browser — nothing is uploaded to any server, and there is no backend database. The project is open-source on GitHub; you can audit the code or self-host.

We recommend staying under 380 characters — this approximates the token limit of Stable Diffusion's default CLIP text encoder, beyond which tokens may be truncated or weakened. The tool shows a live character counter. Midjourney, GPT-image-2, and Nano Banana are more lenient on length, but concise prompts generally produce better results.

Positive prompts describe what should appear (e.g. masterpiece, best quality, detailed face); negative prompts describe what to avoid (e.g. lowres, blurry, bad anatomy, extra fingers). The "Negative" button copies a curated set of common negative prompts to your clipboard — paste it into your model's Negative Prompt field.

The right panel includes a Translate button: enter text in any language, click Translate, and the tool calls a translation API to convert it to English and append it to the prompt box. Handy for starting from an inspirational sentence instead of looking up each word.

In Stable Diffusion WebUI syntax, (word) raises the weight to 1.1, ((word)) to 1.21, and so on; you can also write (word:1.3) to set it explicitly. Midjourney uses ::2 syntax for weights. IMGPrompt preserves the original English tokens so you can fine-tune weights per your target model's rules.

IMGPrompt is a click-to-build prompt composer, not a prompt search engine. PromptHero, Lexica, and PromptBase let you browse or buy other people's prompts; IMGPrompt lets you compose your own by clicking bilingual visual tags. It runs entirely in your browser with no login and no upload, while those platforms require accounts and track your activity.

Yes. IMGPrompt builds prompts for video models including Google Veo, Kling, Seedance, Runway, and Hailuo, alongside image models. Most prompt sites focus on image models only, so IMGPrompt is one of the few free tools that covers the new wave of AI video generation.

Yes. The UI is available in 18 languages, and every tag shows both English and your local-language label. You click tags in your language; IMGPrompt assembles the prompt in English (which AI image and video models actually need). The Translate button also converts any sentence in your language into an English prompt in one click.