Get Inspired: Writing Midjourney Prompts with ChatGPT

In this newsletter, read about:

• 🕵️‍♀️ Midjourney + ChatGPT

• 🗞 News and Top Reads

• 📌 AI Art Tutorial: Midjourney V5 + ChatGPT-4

• 🎨 Featured Artist: IVA

• 🖼 AI-Assisted Artwork of the Week

• 🤓 How to Get Started with Generative AI?

🕵️‍♀️ Midjourney + ChatGPT

As you probably already know, from this newsletter or elsewhere, Midjourney V5 works best with prompts written in natural language. And who else is great at natural language? Right, ChatGPT! 😆 

So, let’s see if we can get more creativity and efficiency from a combination of these tools.

Creating Prompt Structure

I like to follow a specific structure in my Midjourney prompts. This allows me to include all the details and get more controllable results. But obviously, this is not required – you may get truly amazing results with messy prompts, where you just list all you want to see in the generated image.

However, if you are like me, and want to add some structure to your Midjourney prompts, here’s a basic format that I follow.

I start by asking (and answering) a few questions:

1. What type of image do I want? Is it a photo, a painting, or an illustration?

2. What’s the subject? For example, is it a person, an animal, or a house? If person, what kind of person? What’s their clothing?

3. What are some other features that I want to add? Do I want to specify background information, lighting, color scheme, or point of view?

4. Do I want the image to reflect a specific style? For example, should it be impressionistic, pop art, or night photography?

5. Do I need the image to be horizontal, vertical, or square? Which aspect ratio do I want to specify at the end of my prompt? E.g., --ar 2:3 or --ar 16:9.

Then, basically the format looks like this:

[Type] of [subject] with [features] in [style] --ar [ar]  

Here are a few examples, kindly suggested by ChatGPT:

1. A hyperrealistic photo of a serene lakeside scene with lush greenery and soft morning light in the style of landscape photography --ar 16:9.

2. A 3D render of a futuristic cityscape with flying cars and towering skyscrapers in the style of cyberpunk --ar 21:9.

3. A photo of a vintage steam train chugging through a mountain landscape with dramatic clouds in the background in the style of black and white photography --ar 4:3.

4. A painting of a medieval knight on horseback with a majestic castle in the background in the style of romanticism --ar 3:4.

5. An illustration of an astronaut floating in space with Earth and distant stars in the background in the style of retro-futurism --ar 1:1.

Let’s see how you can train ChatGPT to create these basic prompts. After mastering the basics, you’ll be able to adjust the training process to create more advanced prompts for your specific needs.

Training ChatGPT as a Midjourney Prompt Generator

I have been using ChatGPT for a while, but this was my very first experiment with training it to generate prompts for Midjourney. So here’s how I started our conversation.

ChatGPT replied positively and even included a few examples in the requested format:

The Midjourney prompt generator is ready for experiments! Now, how do we use it?

Leveraging ChatGPT “Skills”

First of all, we need to acknowledge that ChatGPT is not a miracle tool, but a machine learning model trained on a huge text database to generate the most probable responses, given certain context (note: this is a very simplified explanation).

So, it’s very far from human intelligence, but it has a much (MUCH!) larger vocabulary than any person and can generate multiple different ideas in a matter of seconds. Hence, AI creators have several possible use cases for ChatGPT in the context of image generation:

1. You want to generate an image of a certain subject, but don’t know, which specific details to add for a more interesting image.

   ChatGPT may suggest you multiple different dragon descriptions with lots of details. You are free to choose!

2. You’ve tried multiple different prompts, but cannot get the desired output.

   Multiple users who encountered similar issues, report that they were able to finally get the required output after ChatGPT reworded their prompt.

3. You want to create a photo in a style of a specific photographer or artist.

   Instead, or in addition, to using their name, you can get ChatGPT’s help to include some distinguished characteristics of their work, like camera type, focus, depth of field, etc.

4. You want to create a series of images, where different subjects will be depicted in the same style.

   ChatGPT can automatically create for you dozens of prompts that would only differ in the subject details.

For example, I had a series in mind, where I wanted to create photo portraits of women from different nationalities, but depicted in a similar style. So, I have continued my conversation with ChatGPT.

After a while, I recognized that I can improve my experience if generated prompts would include “/imagine prompt: ” right away.

As you can see, ChatGPT can be very flexible and you can totally adjust it to your needs in image generation.

Now let’s see some results!

Enjoying the Results

Below you’ll see Midjourney images, generated with prompts by ChatGPT, trained by the author. A human is still in the loop, but AI is definitely taking over 😂 

I’m currently working on a Midjourney Prompt Guide, where I plan to share 100+ prompts and images across different subjects and styles.

If you are not yet subscribed to this newsletter, make sure to subscribe as I will share this guide with my subscribers for free as soon as it’s ready.

Happy prompting!

🗞 News and Top Reads

• Stability AI has recently released Stable Diffusion XL – or SDXL for short. This upgrade marks a significant milestone in the evolution of Stable Diffusion.

  • One of the key benefits of this upgrade is the ability to create images using shorter prompts and generate words within the images.

  • Moreover, the quality of the output has been enhanced to a level comparable to Midjourney.

  • The beta version of SDXL is available for testing on DreamStudio and other leading imaging applications.

• ConrolNet 1.1 has been officially released last week.

  • ControlNet is a neural network structure to control diffusion models by adding extra conditions.

  • The new version includes all previous models with improved robustness and result quality. Several new models are added.

  • The update for the AUTOMATIC1111's Stable Diffusion web UI is in beta testing.

• An AI-generated image won a major photography award:

  • German artist Boris Eldagsen's entry won the creative open category at last week's Sony World Photography Award.

  • The artist has refused his prize after revealing his work was created using AI.

  • "AI images and photography should not compete with each other in an award like this," said Boris Eldagsen.

📌 AI Art Tutorial": Midjourney V5 + ChatGPT-4

In this video tutorial, Samson Vowles shares his approach to using ChatGPT as a prompt generator for Midjourney V5. He shows how ChatGPT can understand your preferences and generate prompts tailored specifically to your interests. You simply provide a few keywords or phrases related to your project, and ChatGPT does the rest.

🎨 Featured Artist: IVA

IVA explores the depths of the AI world. Her digital creations are truly outstanding, captivating, and brimming with wonder, vitality, and artistic flair. Each piece she produces is a testament to her exceptional talent, as well as her ability to infuse technology with humanity and emotion.

🖼 AI-Assisted Artwork of the Week

🤓 How to Get Started with AI Art?

1. DALL-E: Creating Images from Text – introduction to text-to-image generation.

2. The DALL-E 2 Prompt Book – a guidebook by OpenAI that explains how to effectively right prompts to generate images across different domains (e.g., photography, illustration, art history, 3D artwork).

3. Best Midjourney Prompts – a guide that covers the basics of Midjourney prompts (e.g., which keywords to use to create abstract art, surreal art, minimalism, etc) as well as some more advanced options (e.g., keywords related t/o camera lenses and filters, imitating certain artists and photographers without using their names). Finally, they provide a list of 600+ creative text prompts for image generation.

4. Stable Diffusion Prompt Book – a prompt book prepared by OpenArt. The book discusses ideal prompt format, using modifiers to change the style, format, or perspective of the image, applying ”magic words” to improve image quality, adding negative prompts, and adjusting Stable Diffusion parameters.

Share Kiki and Mozart

If you like this newsletter and know somebody who might also like it, feel free to share this newsletter. Let’s have more people learn about AI art!

If you have been forwarded this email and you like it, please subscribe below. And welcome to the world of AI art!