I made my first custom #ChatGPT. While I made it for my own use, you are welcome to try it and adapt it.
(Unfortunately, using my #CustomGPT or creating your own requires a GPT-Plus account.)
https://chat.openai.com/g/g-pOP5JGs63-diverse-image-creator
It’s very simple. From the user’s request and generates four landscape images that are very different from each other. It shows the exact image generation prompts it uses so that you can edit that prompt if you like I use it for my linked cover art.
I no longer have to type in instructional details each time, just information about the particular. It is amazing how big a quality-of-life improvement this is for me.
Once I figure out color palettes and artistic styles that work well for LinkedIn articles, I’ll add that in. Our whole group can have styling that is consistent. I might even have it save the images and prompts in a common area for all of us.
In case you’d like to make your own, you can start with my initial version. Then you can go on to make one useful for your needs.
The complete configuration is:
- checking the “use Dall-E” checkbox,
- uncheck all the others and the instruction text:
And this in the instruction box:
The user will supply a short image description.
The GPT will produce 4 detailed image generation prompts that are very different from each other in as many ways as possible while still remaining true to the user’s image description.
The images should be 1920×1080.
The GPT should show the user the first prompt and the image for it.
The GPT should show the user the second prompt and the image for it.
The GPT should show the user the third prompt and the image for it.
The GPT should show the user the fourth prompt and the image for it.
Yes, it is “idiot simple.”
Yes, the format of its out varies widely from run to run.
But it only took 10 minutes from start to finish to create and has generated images for me in more than 200 instances where I would not otherwise have bothered.
