AI WORLD HQ
You need a picture that doesn't exist yet — a birthday card, a blog header, a poster for your band's next gig. Midjourney is one of the most popular AI image generators (tools that create pictures from text descriptions), and it works well even if you can't draw. Here's how to actually use it.
Before you start
You'll need:
- A computer or phone with a modern browser (Chrome, Safari, Firefox)
- An email address, Google account, or Discord account
- A payment method (credit card or PayPal)
- Around $10 USD for the cheapest plan — Midjourney no longer offers a reliable free tier
Midjourney runs at midjourney.com in your browser. There's also a Discord version, but the website is calmer for first-timers and easier to follow with screenshots.
Step 1 — Create your account
Open midjourney.com in your browser. In the top-right corner, click Sign In. You'll see options to continue with Google, Discord, or Email. Pick whichever you already use — Google is fastest for most people. After you authorize the sign-in, Midjourney redirects you back to its home page.
💬 Example path: midjourney.com → top-right Sign In → Continue with Google
You'll know it worked when: you see your username or profile picture in the top-right corner of the page.
Step 2 — Pick a subscription plan
Midjourney will prompt you to subscribe before you can generate anything. Look for a Subscribe or Get Started button — it's usually visible on the home page or in the left sidebar. The three main plans are Basic ($10/mo), Standard ($30/mo), and Pro ($60/mo). Basic gives you about 200 images per month, which is plenty while you're learning. Pick a plan, enter your card details, and confirm.
💬 Example: Choose Basic Plan → Monthly billing → Confirm payment.
You'll know it worked when: your account shows an active subscription, and the prompt input box becomes active (no longer greyed out or locked).
Step 3 — Write your first prompt
On the Midjourney home page, find the text input area — usually labeled with words like "Imagine" or "Describe what you want to create". Click inside the box and type a description of the image you want. Keep your first one simple. Then press Enter or click Generate.
💬 Example prompt to copy:
a cozy coffee shop at golden hour, watercolor style, soft warm light
After about 30–60 seconds, you'll see a grid of four image options appear below your prompt.
You'll know it worked when: four finished images appear on screen within a minute or two, and no error message shows.
Step 4 — Pick a favorite and try variations
Below each of the four images, you'll see two rows of buttons: U1–U4 (upscale that image) and V1–V4 (create variations of that image). Hover over them to see the full label — usually Vary (Subtle) or Vary (Strong). Click Vary (Subtle) on whichever image you like best to get four close cousins of it. Try Vary (Strong) if you want wilder, more different alternatives.
💬 Example: If image #2 is your favorite, click V2 → Vary (Subtle) to refine it.
You'll know it worked when: a new grid of four refined images appears within a minute, based on the one you picked.
Step 5 — Upscale your final choice
When you find an image you love, click the matching U button (U1–U4). Midjourney processes the image for a few seconds and shows it at full resolution with sharper detail. On the web interface, you can also click the image itself to open a larger preview.
💬 Example: Click U3 to upscale the third image of your latest grid at full quality.
You'll know it worked when: a single, larger image appears (typically 1024×1024 pixels or more), with no "loading" spinner.
Step 6 — Save your image
Click your final image to open it in Midjourney's preview window. Look for a download icon (usually a small downward arrow) in the top-right corner, or right-click the image and choose Save image as…. Save it to your Downloads folder, and you're done.
💬 Example path: Click image → top-right download arrow → Save as "my-first-midjourney.png"
You'll know it worked when: the image file appears in your Downloads folder and opens normally on your computer or phone.
Common mistakes
- Writing a prompt that's too vague. "A nice picture" gives you generic results. Add specifics: subject, setting, mood, style. "A cat on a windowsill at sunset, oil painting style" beats "cat picture" every time.
- Always picking image #1. Midjourney's strongest output isn't always the first one. Scan all four results — your favorite is often #2 or #3.
- Trying to download before upscaling. Click the U button first. Otherwise you'll save a low-resolution thumbnail, not the full-quality image.
- Pressing Enter with no plan active. If nothing happens when you submit a prompt, you probably skipped Step 2. Subscribe first, then the prompt box unlocks.