I was building out a Facebook blog about dogs and what they can and cannot eat when I realized something was missing: good images. Stock photos are expensive, recycled everywhere, and they kill your credibility when readers have already seen the same picture on twelve other websites. So I started making my own images with Midjourney V5, and what came out genuinely surprised me.
These were AI images of dogs eating peppercini peppers, dogs with Reese’s Puffs cereal, dogs dressed as pirates, and Doberman Pinschers in military uniforms. When I looked at them, I could barely tell they were AI-generated. That is where AI image creation is right now, in 2023, at the V5 level. If you have ever wanted to create your own custom illustrations for a blog, a digital product, or an Etsy shop, this walkthrough is for you.
What You’ll Walk Out With
- A clear picture of what Midjourney V5 actually costs and how to access it
- The exact prompt format that gets you a usable image in under a minute
- Three real-world ways to monetize the images you create
- How to use the U1, U2, U3, U4 buttons to pick and save individual images
- A free hack for getting prompt ideas when your brain is blank (hint: it involves ChatGPT)
- The honest limitations of AI art so you can plan around them
- A path to discovering which AI-powered skill fits your actual situation at finder.platformproof.com
Why Midjourney V5 Is a Different Animal
Midjourney has gone through several versions, and the jump from earlier iterations to V5 is significant. Earlier versions of the tool produced results that were creative but obviously AI-generated: strange anatomy, blurry textures, ears that looked melted. V5 cleared most of those problems up. The outputs are sharp, the compositions hold together, and in many cases a first-time viewer will not notice the image was machine-made at all.
For context, earlier AI-generated images of dogs eating cereal had a noticeable fakeness to them. The version I showed in this video had dogs that looked genuinely photo-real next to their food. That is how much the tool has improved in a short amount of time, and it is the reason this is worth paying attention to right now.
The $20 Per Month Access Question
Midjourney is not free, but it is not expensive either. The basic subscription runs twenty dollars per month. That is the same price as a Netflix subscription or a tank of gas in many cities. For a tool that can generate unlimited custom images across any topic or style you want, twenty dollars a month is a reasonable cost of doing business.
You access Midjourney through Discord, which is a free chat platform. You join the Midjourney server, subscribe through their website, and then you can generate images directly inside Discord channels or in a private message with the Midjourney bot. There is no separate software to install and no complicated setup process. If you can type a sentence, you can use Midjourney.
The Use Case That Started This
Here is the specific project I was working on when I started using Midjourney V5. I was building a Facebook blog around the topic of dogs and what they can and cannot eat. Every post needed a featured image and sometimes additional illustrations. Questions like “Can dogs eat peppercini peppers?” need a matching visual to look professional and trustworthy.
I was using ChatGPT to write the blog posts themselves and Midjourney to generate the images. The combination of the two tools let me produce content without paying for a stock photo subscription, hiring a designer, or learning to illustrate anything myself. The images I generated looked good enough to publish without any editing or post-processing.
That combination of ChatGPT for text and Midjourney for visuals is one of the simplest two-tool stacks in the AI space right now. If you are testing out content-based ways to make money online, this setup removes one of the biggest friction points: finding images that actually fit your content.
The Basic Prompt Format
The way you talk to Midjourney is through a prompt. A prompt is just a short text description of the image you want. You type it into Discord and Midjourney generates four image options in about 30 to 60 seconds depending on server load.
The key thing to know about Midjourney prompts is that you have to start them with the word /imagine followed by a space. That is the command that tells the bot to generate an image. Everything after that space is your actual description.
So a simple prompt looks like this:
/imagine dog with gold chain
That is it. Four words after the command and you get four different images of dogs wearing gold chains. They will not all be the same breed or style or lighting, which is exactly what you want when you are picking the best option for your project.
Three Real Prompts I Tried (And What Came Back)
In the video I ran several live prompts to show what Midjourney V5 actually produces. Here is what I typed and what came back.
Prompt 1: Dog With Gold Chain
This was the first test. I typed /imagine dog with gold chain and Midjourney gave me four different dogs, each wearing a different style of gold chain. The breeds varied, the lighting was different in each frame, and the chain detail was sharp and realistic. If you needed a fun editorial image for a pet-lifestyle blog post, any of these four options would work.
One thing you can do to get more variation is add background or setting details. I mentioned you could add “on a white background” or “in New York City” to the same prompt and get completely different results. Each modifier changes the composition and mood of the output without you needing to understand anything about photography or design.
Prompt 2: Dog as a Pirate
The second test was /imagine dog as a pirate. This one took a minute or two to render and came back with four variations of dogs in pirate costumes. These were portrait-style images with hats, coats, and accessories that you would not be able to tell from a human-made illustration if you glanced at them quickly. They looked like the kind of thing you would see sold as art prints in an online shop.
This is the kind of prompt where you can immediately see commercial potential. Search “dog pirate art” on Etsy and you will find sellers moving canvas prints and digital downloads of exactly this type of image. Midjourney produces this output in under two minutes, for free beyond the $20 subscription cost.
Prompt 3: Doberman Pinscher Dressed as a Military Sergeant
This was the most detailed prompt of the three: /imagine Doberman Pinscher dressed up as a military man. I ran this one twice, once and then again after hitting the regenerate option, to show that you can ask for entirely new variations if you do not like the first set of four.
What came back was a series of serious, portrait-style images of Dobermans in military uniforms. The breed was accurate, the uniforms had detail, and the overall quality was high. I noted on camera that this was “frighteningly good” because even knowing it was AI-generated, the images held up at a glance. That level of specificity, naming a breed and a specific costume, still works reliably in V5.
How to Pick and Save Your Best Image
Every Midjourney generation gives you four images arranged in a 2×2 grid. Below that grid you will see buttons labeled U1, U2, U3, and U4, along with V1, V2, V3, and V4.
The U buttons stand for Upscale. When you click U1, Midjourney takes the top-left image (image number one) and renders a larger, higher-resolution version of just that image. U2 upscales the top-right image, U3 the bottom-left, and U4 the bottom-right.
The V buttons stand for Variation. Clicking V1 asks Midjourney to generate four new images that are similar to image number one but not identical. This is useful when you are close to what you want but not quite there.
To save an image, you click on the upscaled version to open it full size, then right-click and save it to your computer. From there it is yours to use however you want within Midjourney’s terms of service. The whole process from prompt to saved file takes about two to three minutes.
What to Do When Your Ideas Run Dry
One of the things I said in the video is that the only limits with Midjourney are money and your mind. But sometimes your mind goes blank, especially when you are trying to come up with a lot of image ideas quickly for a content project.
The easy fix is to use ChatGPT as a prompt generator. You can ask it something like “Give me 20 Midjourney prompt ideas for a dog-themed Etsy shop” or “Write 10 detailed Midjourney prompts for an adult coloring book about forest animals” and it will hand you a full list. Then you run those prompts in Midjourney one by one and see what comes out. You are essentially using one AI tool to feed the other, which cuts your research time down to almost nothing.
Three Real Ways to Make Money With These Images
Here are the three monetization paths I mentioned in the video. None of them are guaranteed income and results depend on your effort, niche, and marketing. But these are real business models that people are using right now.
1. Etsy Digital Downloads and Print-on-Demand
Etsy has a large and active market for pet-themed art. Canvas prints, wall art, and digital downloads of dogs, cats, and other animals in funny or creative costumes sell consistently on the platform. I showed this during the video by scrolling through Etsy listings, including a cat dressed up in a military uniform that was selling as a canvas print. Someone made that with AI or a similar tool.
With Midjourney you can generate dozens of variations quickly, list the digital files for immediate download on Etsy, and collect payment with no inventory or shipping required. If you want to sell physical prints, print-on-demand services let you upload your images and they handle printing and fulfillment when someone orders.
2. Amazon KDP Coloring Books
I showed an example during the video of an adult coloring book page: a bear in a forest scene with detailed line work. Midjourney can generate coloring-book-style illustrations by specifying “coloring book style” or “black and white line art” in your prompt. You can generate an entire coloring book’s worth of pages and then format them into a PDF to upload to Amazon Kindle Direct Publishing as a paperback.
KDP coloring books are a real niche. People search Amazon for themed coloring books constantly, whether that is dogs, fantasy animals, floral designs, or anything else. The barrier to entry has dropped significantly now that you can generate the illustrations in minutes instead of hiring an artist or drawing them yourself.
3. Blog and Facebook Content Images
This is the use case that started my own exploration: using AI-generated images to illustrate blog posts. If you are building a content site or a Facebook page around a niche topic, images are a constant need. Stock photo subscriptions can cost $30 to $100 per month and often do not have exactly what you need anyway. With Midjourney at $20 per month, you can generate custom images that match your exact topic every single time.
For a blog about dog nutrition, every post can have a matching hero image of the specific food and dog breed in question. That level of customization was not practically achievable before AI image tools existed unless you had a photographer and a dog on call.
Not sure if AI art is the right path for you, or if there’s a better fit given your skills and schedule?
The Platform Proof Finder matches you to the most realistic first-income method based on your actual situation. Try it free at finder.platformproof.com.
Selling Prompt Writing as a Skill
One angle I brought up in the video that often gets overlooked: prompt writing itself is a marketable skill. As AI image tools become more widely used, more businesses and creators will want good images but will not want to spend time learning how to write effective prompts. If you get good at writing prompts that produce reliable, high-quality results, you can sell that skill.
This could mean writing prompt packs and selling them as digital products, offering a prompt-writing service to other creators, or taking on clients who want a set of custom images generated for their brand or project. The skill is still early enough that most people have not developed it, which means the people who do develop it now have a head start.
Honest Drawbacks of Using Midjourney V5
I want to be straight with you about where this tool is not perfect, because going in with realistic expectations is important.
It is not free. Twenty dollars a month is not a lot, but it is a real cost. If you generate a lot of images and want faster rendering or more monthly generations, higher plans cost more. You need to factor this in against whatever revenue you are hoping to earn from the images.
The outputs are not always exactly what you asked for. Midjourney interprets prompts rather than following them literally. Sometimes the first generation misses the mark and you have to iterate a few times to get something usable. This is part of the learning curve and it gets easier as you practice writing prompts, but expect some trial and error at the start.
You are operating in a competitive space. The barrier to entry is low, which means a lot of people are generating AI images right now. If you plan to sell on Etsy, you will be competing with a growing number of AI-assisted shops. The people who will stand out are those who pick a specific niche, develop a consistent style, and build a real brand around their work rather than just flooding the market with random generations.
Terms of service and copyright are still evolving. Midjourney’s terms of service govern what you can do commercially with the images. These rules have changed as the tool has grown, so it is worth reading the current terms before building a business around this. The situation is different for free plan users versus paid subscribers.
Some outputs still look AI-generated. While V5 is remarkably good, it is not flawless. Hands, text within images, and certain complex scenes still produce noticeable errors in some generations. The dog images I showed held up well, but this is not a tool that will replace professional photography in every situation.
A Simple Starting Plan If You Want to Try This
If you want to test whether AI art can work for your situation without overcommitting, here is a low-risk way to start.
- Sign up for Midjourney at $20 per month and plan to spend the first week just practicing prompts without worrying about monetization yet
- Pick one narrow niche you actually find interesting, whether that is dogs, cats, fantasy creatures, architecture, food, or anything else with a clear Etsy or KDP audience
- Generate at least 50 images in that niche using variations of the same core concept so you start to see what the tool does well and where it struggles
- Select your best 10 images and list them as digital downloads on Etsy to test whether there is any market interest before you invest more time
- If you see any sales or significant traffic in the first 30 days, double down on that niche and expand the product line; if not, pivot to a different niche or monetization path
This approach limits your downside to the cost of one month of Midjourney plus the time you spend generating images. If it does not work in the first month, you have spent $20 and gained experience with a tool that is useful in other contexts too.
Find Your X
AI art is one path, but it is not the only one. The right first-income method depends on your available time, your budget, your existing skills, and how comfortable you are with creative work. If Midjourney feels like a fit, the steps above are a solid place to start. If you are not sure, or if you want to compare this path against other options before spending money on tools, the Platform Proof Finder was built to help you figure that out. It is free and takes about two minutes to complete at finder.platformproof.com.
Frequently Asked Questions
Do I need any design experience to use Midjourney V5?
No. The only skill you need is the ability to write a short text description of what you want. You do not need to know anything about composition, color theory, or graphic design. The tool handles the visual work based on your written instructions.
How long does it take to generate an image?
Most generations take between 30 seconds and two minutes depending on server load and the complexity of your prompt. Upscaling an individual image from the initial grid of four adds another 30 to 60 seconds. You can run multiple prompts in parallel to speed things up.
Can I sell the images I make with Midjourney?
Paid subscribers can use their images commercially under Midjourney’s current terms of service. Free plan users have more restrictions. This is an area where the rules have changed over time, so checking the current terms on the Midjourney website before selling is important. Do not rely on older guides or forum posts since the policy has evolved.
What is the difference between the U and V buttons?
The U buttons (U1 through U4) upscale one of the four generated images into a larger, higher-resolution file. The V buttons (V1 through V4) generate four new images that are variations of the selected image. Use U when you have found an image you like and want a final version to save and use. Use V when you are close but want to see more options that look similar.
Is Midjourney the only AI image tool worth using?
No. Dall-E from OpenAI and Stable Diffusion are two other widely used options. Midjourney V5 produces some of the best results for photorealistic and artistic images at its price point, which is why I use it for this type of work. Other tools have their own strengths and some are free, so it is worth knowing they exist if the $20 per month is a barrier.
Can I use ChatGPT to help write better Midjourney prompts?
Yes, and this is one of the most practical combinations you can set up. Ask ChatGPT to generate a list of Midjourney prompts for your specific niche and it will give you a starting point you can run directly in Discord. You can also ask it to add style descriptors or camera settings to make your prompts more specific and get better results.
What niches work best for selling AI art on Etsy?
Pet-themed art (dogs, cats, rabbits) has a proven audience on Etsy. Fantasy animals, botanical illustrations, and character portraits also sell well. Niche-down works better than general: “golden retriever as a knight” is more specific and more searchable than “dog art.” Adult coloring books and printable wall art are two formats that move consistently as digital downloads on the platform.
How do I make sure my AI images look good enough to sell?
Always upscale the image using the U buttons before saving. Use V5 (or the current best version) rather than older versions since the output quality improves significantly between generations. Write more specific prompts rather than generic ones, specifying breed, outfit, setting, and lighting style when relevant. And generate a lot of options before committing to a listing: running ten or twenty prompts and picking the best three gives you a much stronger product than listing the first thing that comes out.
Read Next
If you want to explore how AI-generated images fit into a broader digital product business, the next logical step is understanding what kinds of digital products actually sell and how to package them for consistent income.
This post walks through the whole picture: The Easiest Digital Product to Make and Sell in 2026 (Even With No Audience).
Sources
- Midjourney V5 official tool and pricing: midjourney.com
- Etsy pet art market: etsy.com (search results for dog canvas and dog wall art)
- Amazon KDP coloring book publishing: kdp.amazon.com
- ChatGPT prompt generation: chat.openai.com
- Midjourney terms of service: docs.midjourney.com/docs/terms-of-service
Helping 1 million working adults make their first $3,000 online with the skills they already have. Alston Godbolt, Platform Proof.