Updated October 7, 2026 for Midjourney V8.2. We checked the steps against Midjourney’s official documentation and update notes.
Midjourney turns a short written description into images. This guide shows how to start, what each step does and the settings worth learning first.
What you need
- A Midjourney subscription. Midjourney’s docs list four plans: Basic, Standard, Pro and Mega. They say there is no free trial on the website or in Discord (a limited trial exists on the niji journey mobile app). Check the plans page for current prices.
- A browser. You can create on midjourney.com. Discord is optional: Midjourney also supports it (see its Discord Quick Start in the docs).
Step 1: Open the Create page
Go to midjourney.com and open the Create page. At the top is the Imagine bar, where you type your prompt.
Step 2: Write a simple prompt
Type what you want to see and press Enter. Midjourney generates a set of four images. Midjourney’s own advice is that short, simple prompts usually give the best images. A prompt can be one word or a full phrase.
A good starting formula is subject, setting, light and style:
a lighthouse on a basalt coast at blue hour, long exposure, soft sea hazeStep 3: Pick, vary and refine
Open an image to see it larger. From there you can create variations, upscale it, or open the Edit tools. Change one thing at a time so you can see what each word does. See how to write better AI image prompts.
Step 4: Learn the first parameters
Parameters are short settings added to the end of a prompt. The most useful early ones, from Midjourney’s parameter list:
| Parameter | What it does | Values |
|---|---|---|
--ar | Aspect ratio (shape of the image) | Width:height, such as --ar 16:9; default is 1:1 |
--stylize or --s | How strongly Midjourney’s own style is applied | 0 to 1000; default 100 |
--chaos or --c | How varied the four results are | 0 to 100; default 0 |
--weird or --w | Adds experimental, unusual results | Up to 3000; default 0 |
--no | Leaves something out | --no clutter, people |
--seed | Reuses a starting point for tests | A number |
--v | Chooses the model version | V8.2 is the default |
--p | Applies your personalization profile | No value needed |
--draft | Fast, low-cost draft images | See below |
Example with parameters:
a lighthouse on a basalt coast at blue hour, long exposure, soft sea haze --ar 3:2 --s 250See Midjourney parameters explained for examples.
What is new in V8.2
- V8.2 is the default model since July 24, 2026, focused on aesthetics, image quality and personalization. Personalization uses your ratings, so rate images to improve it. Your V7 global profile works with V8.2.
- The Edit Model (August 27, 2026) edits images with instructions, uses up to four image references (replacing Omni Reference and Character Reference) and supports inpainting and outpainting. Drag images into the prompt bar under “attach to prompt”, click “edit” in an image’s lightbox, or use the Edit tab.
- Draft mode (
--draft) is for quick experiments. In V8.1 and V8.2 a single draft prompt makes a batch of 24 lower-resolution images (512 by 512) and costs 0.4 minutes of GPU time. - Fast and Relax modes: your plan includes a set amount of Fast GPU time. Relax mode, with unlimited slower generation, is available on Standard, Pro and Mega.
Common beginner mistakes
- Writing prompts that are too long. Start simple and add detail gradually.
- Changing many things at once. Change one word at a time.
- Ignoring aspect ratio. Decide the shape first with
--ar. - Skipping personalization. Rating images makes results closer to your taste.
- Using old tutorials. Many guides still describe V5 or V6 settings; check the version.
Ready for more?
- Midjourney parameters explained
- Advanced techniques for mastering Midjourney
- Midjourney landscape prompts and portrait prompts
- Midjourney review and how it compares with Flux and GPT Image
