Midjourney prompts with the resulting image attached to each, so you can pick by outcome rather than guess from the text. Copy one, paste it into Discord or the Midjourney web app, and tune the parameters from there.
Moody black and white street photography of a romantic couple sharing a single black umbrella under heavy rain on a cobb...
Gritty cyberpunk street photography of a male techwear model standing on a rain-drenched rooftop in Neo-Tokyo. Wearing a...
Transform the person into a traditional Navratri celebration scene. They should be wearing vibrant traditional Indian at...
Generate an image from the uploaded photo that reimagines the subject as an ultra-detailed 3D graphite pencil sketch on ...
Create a photograph of the person in this image as if they were living in the 1970. The photograph should capture the di...
An Indian man with a similar face structure, full beard, and styled hair is leaning casually against a vintage white Amb...
A stunning and glamorous woman with long, flowing dark hair and a confident gaze looks directly at the camera. She is we...
Explore the full library — 1,000+ prompts, all free to copy.
Use the web app at midjourney.com or the Discord bot, whichever you already have access to.
In Discord, type /imagine and paste the prompt after it.
Add --ar 4:5 for portrait or --ar 16:9 for wide. This changes the composition, not just the crop.
Use U1–U4 to upscale the result you like, or V1–V4 to generate variations of it.
Midjourney weights the front of a prompt most heavily, so the subject belongs in the first few words and the styling behind it. It also applies a strong aesthetic of its own — it will make a plain description look good without being asked, which is a help on concept art and a hindrance on documentary or product work, where you have to explicitly ask for plainness.
Parameters go at the end of the prompt. --ar sets aspect ratio (--ar 3:2 for landscape, --ar 4:5 for portrait) and is the one to reach for first, because framing changes composition more than any adjective. --stylize controls how much of Midjourney's own taste gets applied; lower values follow your wording more literally. --no removes an element. --chaos raises the variation between the four results in a grid. Everything else is refinement.
Reusing the same style clause across prompts gets you most of the way to a consistent set — the same film stock, the same lighting, the same lens. For tighter consistency, use a style reference with --sref and keep the reference fixed while the subject changes. Describing the same character in words alone will not hold their face steady between generations.
The prompts in this library are free. Midjourney itself is a paid subscription — it no longer offers an open free trial, so you will need a plan to run them.
--ar sets the aspect ratio, for example --ar 16:9 for widescreen or --ar 4:5 for a portrait crop. It affects how the scene is composed, not just how the image is cropped, so it is worth setting deliberately.
Yes, as long as you strip the parameter flags. The descriptive text transfers cleanly; anything beginning with -- is Midjourney-specific and other models will read it as literal text.
Use a character or style reference rather than description alone. Midjourney's --cref and --sref parameters hold a face or a look steady across generations in a way that wording cannot.