How to convert image to prompt?

There are many ways to get a tool to describe your image, and extract the “prompts” behind it.

However, you need to get the BEST description that will fit the model that you’re using.

There are three most popular image models on the market right now, Midjourney, Stable Diffusion and now Dalle-3.

We outline the tools and techniques on how to get the best prompt from your image, so that you can recreate it.

Stable Diffusion Image to Prompt

If you want to re-create a specific image in Stable Diffusion, a good tool to extract the prompt is Img2Prompt. This tool is optimized to generate an approximate text prompt with a specific style for Stable Diffusion.

It will give you all the “hacky” details that Stable Diffusion needs in the prompt, in order to generate a similar photo.

You can also integrate the tool directly into your app, from it’s hosted API on Replicate.

MidJourney Image to Prompt

If you want to re-create images in MidJourney, you can use their Midjourney /describe option that will write 3 different prompts (descriptions) based on the image you feed in.

Then using one of those prompts you can re-create a similar image.

You can also get help with your prompts by using tools like the Midjourney Prompt Generator, that will enhance your prompts and give you more ideas to try.

DALLE3 Image to Prompt

There is no official model or tool to extract the “prompt” behind a photo if you want to re-create it in DALLE3. But DALLE3 likes simple, clear prompts. Just describe your photo in as much detail as you can and it should work great.

Another option is to feed the image into Bing Chat, and ask it to generate

If you still need ideas, try using the Midjourney /describe or Stable Diffusion img2text tools.

They're not made for DALLE3 but can give you good hints on how to talk about a photo to a machine.