After the recent update of DALLE3, I mainly use DALLE (inside ChatGPT) to create my images. Before that I was mostly using Midjourney.

So, let’s cut the crap and head over to the step-by-step tutorial!

(You must have a ChatGPT+ subscription to follow this guide to it’s fullest, but you could probably get good results with other image generating tools as well).

1. Go to ChatGPT and upload an image of a picture, character or anything you want to create more of. For me, that is obviously my little SamurAI.

2. Press on the “attach” button to upload your reference picture. I’m using my original picture of my SamurAI character.

3. Enter following prompt into ChatGPT:

   1. “Could you please describe the character in this photo in details, so I can recreate it later with DALLE 3?”

   

4. Now wait for ChatGPT to respond with its description.

   

5. Next, we need to craft our prompt which we’re going to use to generate the actual image we want. Let’s say I want to create my SamurAI sitting working on a computer with YouTube on the screen.

6. I copy-paste the description given to me from ChatGPT Vision, and then edit and/or add the information about the setting for how he is sitting/standing and what he’s doing.

7. The prompt does now look like this:

   “a stylized chibi-like character, reminiscent of a futuristic samurai with robotic head. Here are the details:

   Character's Head:

   The character's head is oversized in proportion to its body, typical of a chibi design. It wears a smooth, rounded helmet that is dark in color, black. The helmet has intricate golden details and patterns adorning it. The most striking feature is the character's eyes: two large, glowing, circular yellow or amber lights, emanating a soft glow.

   Body and Armor:

   The character's body is clad in a detailed armor that shares the same dark color as the helmet, complemented with golden accents and designs. The armor is segmented and appears advanced or futuristic, with sleek lines and ornate golden patterns. There's a noticeable belt with golden embellishments and a buckle.

   Weapons and Pose:

   On his back, the character has a samurai sword, visible over one of his shoulders.

   Ground and Effects:

   The character sits on his chair working on a computer. Big computer screen showing YouTube logo and some graphs on it. Image has black and yellow theme.”

8. You can see at the end of last prompt, that I’ve added how he’s sitting working on a computer with the screen showing YouTube. We now Hit enter and wait for the results.

   

   

9. As you can see, the result wasn’t very good at my first try, but that doesn’t matter much, we just try again. I gave some extra instructions to ChatGPT for it to understand more what I was looking for: “two more please. picture taken behind character, and computer screen must be visible.”

   

10. At my second try, I got an image I was very satisfied with. But I tried once more, to see if it could make it even better. (Prompt: “Thanks, two more please, in the same style as the first image in your last reply”)

11. And in my opinion, it did! :

    

That’s all folks!