AI WORLD HQCreate Your First AI Image with Midjourney: A Beginner Walkthrough
From sign-up to a finished picture saved on your device in one short sitting — no art skills or design background needed.
In short: Midjourney turns a written description into a picture in roughly a minute. You'll need a Midjourney subscription, about ten minutes, and a clear idea of what you want to see. The finished image is yours to download and use.
In the next ten minutes, you'll go from a blank screen to a finished picture saved on your device. No art background needed — if you can describe what you want, you can do this. We'll skip the jargon and focus on the buttons you'll actually press.
💡 Tip: tap a step’s number when you finish it — a green tick appears and your browser remembers how far you got.
- A Midjourney account — sign up on the official Midjourney website using your email, a Google account, or a Discord login. Generating images requires a paid subscription; current plans and pricing are listed on Midjourney's official pricing page, so check there before you commit.
- A web browser or the Midjourney web app — you can work on a computer, tablet, or phone. Discord also works as an alternative entry point if you already have an account there.
- About 10–15 minutes of your time, plus another minute or so while the AI draws.
- One clear idea of a picture you'd like to see. "A cat" works as a starter; "a calico cat asleep on a stack of books, soft morning light" gives a much better result.
Create your Midjourney account
Open your web browser and go to the official Midjourney homepage. Look for a sign-in or sign-up button — it's usually in the top-right corner, often labeled something like "Sign In" or "Get Started". Click it, then sign in with your email, a Google account, or a Discord login.
After signing in, the page will guide you through choosing a subscription plan. Pick whichever option feels right for a first try — you can change or cancel later. Once payment is confirmed, the page reloads and drops you into your personal Midjourney workspace.
If you'd rather start in Discord, the same tool runs inside a chat window there. Either path leads to the same image-quality results, so don't worry about choosing wrong.
You'll know it worked when you see the Midjourney workspace, with a place to type your picture description and your account name visible somewhere on screen.
Find the create screen
Once you're in your workspace, look for the area where you can start a new image. On the Midjourney web app, this is usually a large text box near the middle of the screen, sometimes paired with a sidebar that lists past images or quick-start options. On mobile, the input area may fill the whole screen.
Click anywhere inside that text box. Your cursor should start blinking there. Don't type yet — just confirm you've found it. This box is where you'll describe your picture, so it's worth a moment to make sure you're in the right place.
If your version looks more like a Discord chat, type /imagine in any message field and Discord will offer to autocomplete it. Either way, you're heading to the same destination: a box waiting for your words.
If you don't see a text box at all, look for a button that says something like "Create" or "Imagine," or shows a + symbol — clicking it usually opens the input area.
You'll know it worked when your cursor is blinking inside an empty text box and no error message is showing.
Write your first prompt
Now describe one picture in plain words. Imagine texting a friend who paints for you — tell them the subject, the mood, the setting, and the style. A prompt is just the name for this description; nothing fancy.
Be specific. "A flower" gives you a generic flower. "A single red poppy in a glass jar, morning sunlight, photo-realistic" gives you something that actually looks the way you pictured it. Short sentences are fine. So is a comma-separated list.
When your description feels right, look for a button nearby — usually a paper-plane icon, an arrow, or the word "Generate" — and click it once. Don't click twice; each click spends a credit and starts a fresh image, so patience saves you money.
You'll know it worked when the button greys out, a small progress message appears, and within about a minute the first signs of your picture start forming on screen.
Read the image grid
When generation finishes, you'll usually see a 2×2 grid of four images — four different attempts at your prompt. They're all based on the same description, but each one took the words a slightly different way.
Look at all four. Which feels closest to what you imagined? Which has a color, angle, or detail you really like? You don't have to pick yet — for now, just notice that the four pictures are different from each other, even though you only typed one prompt.
If none of the four feels right, look for a button like "Re-roll" or a circular ↻ arrow somewhere near the grid. That tells Midjourney to try again with the same prompt and produce four new attempts.
You'll know it worked when you can clearly see four distinct images and at least one of them makes you think "ooh, that's nice."
Upscale your favorite
Near the grid, you'll typically find a row of small buttons labeled something like U1, U2, U3, U4 — one under each image. The U stands for upscale, which simply means turning the small preview into a full-size, sharp picture. There are usually also V1–V4 buttons, which create variations on that specific image.
Click the U button under the picture you like most. Within a few seconds, a single full-resolution version of that image appears. This is the one you'll keep.
If the upscaled image has a small detail you'd like changed — say the cat's eyes are the wrong color — you can run a new variation on it, or simply upscale a different image from the grid instead.
You'll know it worked when a single, larger, sharper image appears alongside or in place of the small grid preview.
Download the image to your device
Click the upscaled image to open it full-size. Look for a download control — this might be a small downward arrow ⬇, a button that says something like "Save" or "Download," or an option inside a right-click or menu. One tap or click should save a PNG or JPG file straight to your device.
On a phone, the image often opens in a new tab. From there, look in your browser's menu (often three dots or a share icon) for an option to save the picture to your Photos or Gallery app. On a computer, the file usually lands in your Downloads folder.
The file is now yours. You can post it, print it, set it as your wallpaper, or send it to a friend. Midjourney's standard paid plans deliver images without a watermark, so the picture is ready to share as soon as it's saved.
You'll know it worked when the file appears in your Downloads folder (computer) or Photos / Gallery app (phone), and you can open it again from there.
- Writing a one-word prompt. "Cat" gives you a random cat. Always include at least one detail about mood, style, or setting — Step 3 has an example you can copy and tweak.
- Clicking Generate twice in a hurry. Once is enough. Every extra click spends a credit and starts a fresh image. If the first result isn't ready yet, just wait — a progress indicator should be visible.
- Asking the AI to draw readable text. Midjourney struggles with words inside images — a sign that says "OPEN" often comes out misspelled or jumbled. Use a real text tool or a graphic editor for readable words, and let Midjourney handle the rest of the picture.
Take one minute and write down a single sentence describing a picture you actually want to see. Then open Midjourney, paste it into the prompt box, and hit Generate. By the time you finish reading this paragraph, your first image will probably be ready to look at.
❓ Quick questions
How long does this take?
About 6 minutes — the guide has 6 steps, and you can tick each one off as you go.
Which tool do I need?
This guide uses Midjourney Midjourney — but the approach works very similarly in other AI assistants.
Do I need to prepare anything?
- A Midjourney account — sign up on the official Midjourney website using your email, a Google account, or a Discord login. Generating images requires a paid subscription; current plans and pricing are listed on Midjourney's official pricing page, so check there before you commit.
- A web browser or the Midjourney web app — you can work on a computer, tablet, or phone. Discord also works as an alternative entry point if you already have an account there.
- About 10–15 minutes of your time, plus another minute or so while the AI draws.
- One clear idea of a picture you'd like to see. "A cat" works as a starter; "a calico cat asleep on a stack of books, soft morning light" gives a much better result.
What mistakes should I avoid?
- Writing a one-word prompt. "Cat" gives you a random cat. Always include at least one detail about mood, style, or setting — Step 3 has an example you can copy and tweak.
- Clicking Generate twice in a hurry. Once is enough. Every extra click spends a credit and starts a fresh image. If the first result isn't ready yet, just wait — a progress indicator should be visible.
- Asking the AI to draw readable text. Midjourney struggles with words inside images — a sign that says "OPEN" often comes out misspelled or jumbled. Use a real text tool or a graphic editor for readable words, and let Midjourney handle the rest of the picture.
Keep reading
✦ Original step-by-step guide by AI World HQ's AI editorial team. Written in plain language, reviewed for accuracy.
← Back to all stories