A year ago, getting good results from Midjourney required learning a dense vocabulary of technical parameters and spending hours experimenting. Today the barrier is considerably lower — but the documentation is still scattered enough that most beginners waste their first session confused about where to even start.
This guide skips the history and goes straight to what you actually need to know to get your first useful image out of Midjourney today.
What Midjourney Is and Where It Lives
Unlike most AI tools that run in a browser, Midjourney operates through Discord — a messaging platform originally built for gaming communities. This surprises new users, but the workflow becomes natural within a few minutes.
You’ll need two things before starting: a Discord account (free) and a Midjourney subscription. As of 2025, there’s no free tier — the basic plan starts at $10/month and includes 200 image generations, which is enough to get a genuine feel for the tool before committing further.
Getting Set Up
Once you’ve subscribed at midjourney.com and connected your Discord account, you’ll be invited to Midjourney’s Discord server. From there, you can generate images either in the public newcomer channels or — the better option for beginners — by opening a direct message with the Midjourney Bot and working there privately.
Your First Prompt
Midjourney generates images through a command typed into Discord:
/imagine prompt: [your description here]
The quality of what comes back depends almost entirely on the specificity of your description. Compare these two prompts:
Vague: /imagine prompt: a city at night
Specific: /imagine prompt: aerial view of Tokyo at night, neon reflections on wet streets, cinematic lighting, shot on 35mm film
Both are valid prompts. The second one gives Midjourney a style, a perspective, a mood, and a technical reference — and the output reflects each of those decisions directly.
The Four Images and What to Do With Them
Every prompt generates a grid of four image variations. Below the grid you’ll see two rows of buttons:
U1, U2, U3, U4 — Upscale a specific image to a larger, more detailed version. V1, V2, V3, V4 — Generate four new variations based on that image’s direction.
Most beginners ignore the V buttons and miss half the tool’s value. If one image in the grid has the right composition but the wrong lighting, hitting V on that image generates four new versions that preserve what worked while exploring the elements that didn’t.
Three Prompt Techniques Worth Learning Early
Reference an art style or photographer. Adding «in the style of Studio Ghibli» or «inspired by Ansel Adams photography» shifts the entire aesthetic of the output in a predictable direction.
Specify the aspect ratio. The default is square. Add –ar 16:9 for widescreen or –ar 9:16 for vertical formats suited to social media. This one parameter alone makes outputs immediately more usable for real projects.
Use negative prompting. Add –no [element] to exclude things you don’t want — –no text, –no people, –no blur. Cleaner than trying to describe the absence of something in the main prompt.
Realistic Expectations for the First Session
Your first few prompts will probably underwhelm you. That’s normal — the gap between what you imagined and what generates narrows quickly as you learn which details Midjourney responds to and which it ignores.
Give it an hour of genuine experimentation before forming an opinion. Most people who stick with it past that first session end up integrating it into their workflow permanently.