We’ll show you how to easily generate prompts for Midjourney, Stable Diffusion, and other AI models using existing photos, explore different approaches, and compare two methods.
Let’s say you have a beautiful photo or image, and you want to create a few more similar ones but struggle to describe what’s in the picture. This problem is easily solved using the Image Describer service or good old ChatGPT.
Go to imagedescriber.online and upload the desired photo. You can test the service using sample images. The image size should not exceed 5 MB.
Next, choose one of the suggested prompts, such as Image to Midjourney Prompt, or write your own. You can leave the field blank, and the AI will simply describe what it sees. You can also modify the prompt and ask the AI to describe the image in Russian.
Processing takes just a few seconds, after which you get a prompt that can be used in your preferred AI model to generate new images.
The result differs from the specialized tool but is still great. The chatbot quickly recognized what’s in the image and generated a prompt based on that data. So, if you run out of credits on Image Describer, you can switch to ChatGPT since it has no such limitations.
As a bonus, let’s try generating a prompt with ChatGPT, not from a regular photo but from an unusual artwork. Just like before, we upload the image and enter the request "Generate a Midjourney Prompt for the image." The result is below.