Image Prompt Generator
Genera prompt ottimizzati per DALL-E, Midjourney, Stable Diffusion.
Configura il prompt
Compila il form e clicca "Genera Prompt" per ottenere i prompt ottimizzati per ogni piattaforma.
Come utilizzare Image Prompt Generator
Describe the subject of the image
Describe the main subject of your image, being as specific as possible. Instead of "a cat", write "an adult Sphynx cat sitting on a stone ledge at sunset".
Choose style, atmosphere, framing and lighting
Choose from the menu options for artistic style (photorealistic, anime, watercolor...), atmosphere (dramatic, serene, mysterious...), camera framing and lighting. Each parameter enriches the prompt with technical terminology specific to each AI.
Generate and copy the prompt for your model
Click "Generate Prompt" to get optimized versions for DALL-E 3, Midjourney and Stable Diffusion. Each model has a different syntax: Midjourney uses parameters like "--ar 16:9 --v 6", Stable Diffusion requires negative prompts. Copy your model's prompt with one click.
Suggerimenti
- For Midjourney, always use the parameter "--ar" to set your desired aspect ratio (--ar 16:9 for landscape, --ar 1:1 for social media, --ar 9:16 for smartphone). An incorrect aspect ratio produces images with cut subjects or excessive empty space.
- Mentioning specific artists or art movements in the "Additional Details" field is one of the most effective ways to achieve a consistent style: "in the style of Edward Hopper" or "cinematic composition like Christopher Nolan" convey stylistic information that a list of parameters cannot express.
- Experiment with light and atmosphere combinations: "cinematic lighting + dramatic" yields vastly different results from "golden hour + peaceful". Lighting is often the key to changing the final mood of an image.
Domande frequenti
Why are prompts different for DALL-E, Midjourney and Stable Diffusion?
Every model has a different syntax and sensitivity. Midjourney responds better to descriptive prompts with parameters "--ar" (aspect ratio) and "--v" (version). Stable Diffusion benefits from negative prompts (what you don't want in the image) and specific technical tags like "masterpiece, best quality". DALL-E 3 handles natural language prompts without special tokens.
What are "negative prompts" in Stable Diffusion?
Negative prompts tell the model what to avoid in the generated image. Common terms like "blurry, deformed, ugly, bad anatomy" help exclude typical artifacts of model distributions. Unlike DALL-E or Midjourney, which handle exclusions directly in the positive prompt.
How to improve the quality of generated prompts?
Add specific details in the "additional details" field: materials (red velvet, white marble), precise setting (Venice 1960s, cosmic space), reference artists (in Caravaggio style, inspired by Studio Ghibli) and technical details (macro photography, 8K ultra HD). The more specific details you provide, the more coherent and accurate the generated image will be.
Do generated prompts work with other image AI models?
The DALL-E prompt in natural language is a good starting point for many text-to-image models, but not all support parameters like "--ar" of Midjourney or the negative syntax prompt of Stable Diffusion. Check the documentation of the specific model you use (e.g., Flux, Ideogram) before pasting the prompt like this.
Are my prompts saved or sent to a server?
No. The generation of the three prompts occurs entirely on the client side through string concatenation: no text you type is sent to a server or saved in a database. When you close or reload the page, the form and generated prompts are lost.