The knowledge cut-off for ChatGPT (including GPT-3.5 and GPT-4) is September 2021. This means that GPT is not aware of Midjourney. However, due to how large language models (LLMs) like GPT work, they can be trained with some prompts and produce the desired output.
This means you can teach GPT what you want it to do.
You are PromptGPT. You create detailed prompts for Midjourney, which is an AI image generator that produces images from detailed text prompts. First, you are going to be provided some example prompts. Then you are going to be provided some keywords which you will then use to generate 5 prompts. Before you are provided examples, here is how Midjourney works. - To set the aspect ratio of the image you can use `—ar` to provide an aspect ratio. - Specific camera models, ISO values, f stop and lenses can be used to vary the image produced. - `--chaos <number 0–100>` Change how varied the results will be. Higher values produce more unusual and unexpected generations. - `--Weird <number 0–3000>` Explore unusual aesthetics with the experimental --weird parameter. Prompt examples: /imagine prompt: elderly man, by the sea, portrait photography, sunlight, smooth light, real photography fujifilm superia, full HD, taken on a Canon EOS R5 F1.2 ISO100 35MM --ar 4:3 --s 750 /imagine prompt: film photography portrait of young scottish prince looking at the camera, plate armor, hyperrealistic, late afternoon, overcast lighting, shot on kodak portra 200, film grain, nostalgic mood --ar 4:5 --q 2 /imagine prompt: photograph from 2018s China: a young couple in their 20s, dressed in white, stands in their home, displaying a range of emotions including laughter and tears. Behind them is a backdrop of a cluttered living space filled with white plastic trash bags and torn white paper rolls. Captured with a film camera, Fujifilm, and Kodak rolls, the image conveys a strong cinematic and grainy texture. This artwork uniquely documents the complex emotions and living conditions faced by the young people of that era. --ar 4:3 /imagine prompt: Young, handsome Keanu reeves In a black long leather coat walking down the street in the rain --ar 2:3 —uplight /imagine prompt: flat vector logo of deer head, golden on white /imagine prompt: logo for a jazzy cat cafe with the text: "CATZ" /imagine prompt: rainbows raining down from the sky, cyberpunk aesthetic, futuristic --chaos 50 /imagine prompt: illustration of a dog walker walking many dogs, tech, minimal vector flat --no photo detail realistic Only use the above as examples. Use the following keywords to create new prompts: Dog, t-shirt design, afghan hound
What this prompt does is essentially fine-tunes GPT to produce a desired output. You teach it what you want it to do, provide some additional information and then, in this case, provide some keywords to produce an outcome.
Also, I want to point out that this approach works in ChatGPT but also well with the GPT API. I only tested this with GPT-4, but I imagine it might produce decent results with the updates they have made to GPT-3.5. The upside with GPT-3.5 is the increased context size means you can provide more example prompts and details.
In my instance, I wanted it to produce images that can be used for a t-shirt design of an Afghan Hound dog. The prompts it produced were quite varied and detailed:
/imagine prompt: A detailed illustration of an Afghan Hound, its long silky coat flowing in the wind, standing majestically on a grassy hill. This image is designed to be printed on a t-shirt. The dog is in a playful stance, looking over its shoulder with a cheerful expression. The background is a gradient of sunset colors, creating a warm and inviting atmosphere. The image is rendered in a modern, flat vector style. --ar 1:1 --no photo detail realistic /imagine prompt: A high-resolution photograph of a beautiful Afghan Hound, taken in a professional studio. The dog is positioned in a standing pose, showcasing its unique coat and noble appearance. The image is intended for use in a t-shirt design. Shot on a Nikon D850, f/1.4, ISO 100, 85mm lens, the photograph captures every detail of the dog's features. --ar 4:5 --s 750 /imagine prompt: A hyperrealistic illustration of an Afghan Hound dog, its coat glistening in the sunlight. This design is intended for a t-shirt print. The dog is sitting, looking directly at the viewer with a calm and serene expression. The background is a simple gradient from light to dark, enhancing the focus on the dog. Rendered in Adobe Illustrator for a clean and crisp finish. --ar 1:1 /imagine prompt: A minimalistic, black and white vector illustration of an Afghan Hound for a t-shirt design. The dog is in a dynamic running pose, its coat flowing behind it. The simplicity of the design emphasizes the elegance and speed of the breed. This image is designed with a modern, geometric style, perfect for screen printing on a t-shirt. --ar 1:1 --no photo detail realistic /imagine prompt: An artistic, watercolor-style illustration of an Afghan Hound for a t-shirt design. The dog is lying down, looking up at the viewer with a relaxed expression. The watercolor technique gives the image a soft and dreamy quality, highlighting the dog's gentle nature and beautiful coat. This image is designed in a square format, perfect for a t-shirt print. --ar 1:1
That first prompt produced the following four images (which look great):
The second prompt about an Afghan being shot in a professional studio produced even better results:
The beautiful thing about this approach is you can tailor it to produce images however you want. Want it to produce a logo for your next app idea? Change the keywords.
Logo for time tracking app, blue and green colors
And got the following prompts:
1. /imagine prompt: Logo for a time tracking app named "TimeTide". The design features a stylized hourglass merging with a circular clock face. The top half of the logo is a vibrant blue, representing the fleeting moments, while the bottom half is a calming green, representing the steady progression of time. The logo has a modern, flat design with a slight gradient for depth. --ar 1:1 2. /imagine prompt: Create a logo for a time tracking app called "ChronoCapture". The logo should be a combination of a stopwatch and a camera, symbolizing the capturing of time. Use different shades of blue and green to create a dynamic and eye-catching design. The logo should be in a vector style, flat, and minimalist. --ar 1:1 3. /imagine prompt: Design a logo for a time tracking application named "TimeFlow". The logo should incorporate elements of an hourglass and waves, symbolizing the flow of time. The color scheme should be a gradient of blues and greens, symbolizing reliability and growth. The design should be simple, clean, and modern. --ar 1:1 4. /imagine prompt: Logo for a time tracking app "TimeSage". The logo should incorporate a wise old sage's face subtly morphing into a clock face. The color palette should be a mix of blues and greens, symbolizing tranquility and wisdom. The design should be modern, with a touch of minimalistic style. --ar 1:1 5. /imagine prompt: Create a logo for a time tracking app named "TimeKeeper". The logo should feature a stylized owl holding a pocket watch in its talons. The owl symbolizes wisdom and the watch represents time. Use a color scheme of blues and greens to create a soothing and professional look. The design should be in a flat vector style. --ar 1:1
As you can see, I probably should have provided the name of my fictional time-tracking app. Still, it worked well and hallucinated some app names for me. Also, note that it knew we wanted a logo and set the aspect ratio of our images accordingly.
The first prompt produces some fantastic logo results:
And here is what the second prompt produced:
There is no right or wrong way to do this stuff. The random nature of prompting means you can strike gold with enough effort. As you can see, by “teaching” GPT-4 what you want it to do, it can do things that didn’t exist at its knowledge cut-off.
A special thanks to this other post here (which is well worth the read). I used some of the prompts in that post to guide my example prompts to feed into GPT because it showcases how you can use aspect ratios, camera models and other bits of information to produce different imagery.