Make a square picture of a fluffy orange cartoon puppy, from a blank page to an image saved on your computer. You’ll type a description, choose one of the four puppies Midjourney draws, then refine its face or pose. Everything in the written tutorial happens in a web browser.
1. Open Midjourney and find the drawing box
Go to Midjourney Create. If prompted, sign in to your Midjourney account. Image generation requires an active subscription. If you already subscribe but see a plan-selection screen, check that you signed in with the account attached to your subscription.
On the Create page, look at the long text box at the top. Midjourney calls it the Imagine bar. A prompt is simply the description you type there. This puppy starts from words alone, so you can leave the image-upload area empty.
2. Choose the model and picture shape
Click the settings icon in the Imagine bar. A model is the version of Midjourney that draws the picture. For this fluffy 3D cartoon, use these starting choices:
- Version — V8.2. Select the regular Midjourney V8.2 model. If V8.2 is already selected, leave it there.
- Image shape — Square (1:1). This makes the picture as wide as it is tall. The prompt will tell Midjourney to leave space around the puppy.
- Other controls — leave them at their defaults. For the first set, skip reference images and custom style profiles. You can judge the effect of your words more easily with fewer settings changed.
3. Copy this prompt and generate four puppies
Return to the Imagine bar. Copy the whole line below, including the two settings at the end, and paste it into the box.
A fluffy orange puppy standing on all four paws, round face, large dark eyes, one floppy ear, head tilted slightly to the left, soft 3D cartoon fur, full body centered on a pale blue studio floor, empty space above the ears and below the paws --ar 1:1 --v 8.2
--ar 1:1 asks for a square canvas. --v 8.2 selects the model for this picture. Including both in the prompt also lets you follow the tutorial without changing your usual account settings.
Press Enter once, or tap the send button on mobile. Watch the Create feed below the bar. Wait until the four images finish appearing before choosing one. Your puppy may have different fur or a different expression from the one in the video.
4. Pick one puppy and improve it
Click the image you like best to open it. Check three things at full size: both eyes look in the same direction, the ears fit inside the frame, and the four paws are easy to distinguish. If you like the result, go straight to the download step.
- Good pose, awkward expression? Under Creation Actions, choose Vary → Subtle. This generates close alternatives. Open a new result to compare its eyes and mouth with the first puppy.
- Want a bigger change? Choose Vary → Strong to explore a different interpretation. If Vary is hidden, open More options and enable it.
- Every puppy has cropped paws? Paste the prompt again. Replace “full body centered” with “wide shot from a few steps back, entire puppy visible, generous space around the body.” Keep the orange fur and blue floor so you can judge the framing change.
5. Download the finished image
Open Organize, select the finished puppy, and choose Download. On a computer, check the browser’s downloads list and open the saved file. Zoom in on the face, then zoom out to the size you plan to use. Keep a copy of the prompt with the file so you can make a second pose later.
Watch the original Mr. Technie video tutorial
The video demonstrates the puppy in Discord. When the recording shows /imagine, that is where the description was entered. On the website, the Imagine bar serves the same purpose. The screenshot below shows the puppy chosen in the recording.





























