Free Midjourney Prompts

Public creator results labelled as Midjourney are paired with their source posts and shared prompts — including parameter strings such as --ar, --sref and --stylize when provided. Copy the whole line as a starting point for Discord or midjourney.com.

Use the examples as prompt-writing references rather than guaranteed recipes. Each card shows a published result and the text shared with it; source context, model version, reference images and later edits can all affect what another run produces.

The collection includes portraits, posters, logos, illustration and other public examples. Open a card to read the complete prompt, inspect its model label and follow the source before adapting it.

Browse all prompts →

Recent Midjourney prompt examples

Prompt share: Beautiful women walking down a flight os stairs in a surreal… — Midjourney prompt by @JasonHeatony7
@JasonHeatony7
Midjourney V8.2 : Watercolor portrait illustration. A 20-year-old seductive… — Midjourney prompt by @VIBEQUIRKLABS
@VIBEQUIRKLABS
New Midjourney Style Reference Unlock with --sref 2219508089 Goblin --ar 3:4… — Midjourney prompt by @michaelrabone
@michaelrabone
** A high-fashion studio portrait of a young East Asian woman with long, glossy… — Midjourney prompt by @Monaqbuv
@Monaqbuv
Midjourney V8.2 : Two 20-year-old adult Japanese women. Both women smile gently… — Midjourney prompt by @VIBEQUIRKLABS
@VIBEQUIRKLABS
New Midjourney Style Reference Unlock with --sref 2868354353 Sushi Monster --ar… — Midjourney prompt by @michaelrabone
@michaelrabone
Midjourney V8.2 : Photorealistic editorial portrait. Front view. A 20-year-old… — Midjourney prompt by @VIBEQUIRKLABS
@VIBEQUIRKLABS
Mixing 2 images as styles: 1. Modern pop art geometric abstraction 2. Street… — Midjourney prompt by @michaelrabone
@michaelrabone
Esoteric Don't Look + 35mm Film Midjourney Style + Profile Mix Don't look at… — Midjourney prompt by @michaelrabone
@michaelrabone
Revisiting My Midjourney V6 Styles Here are 4 favourites (more soon) Style 1… — Midjourney prompt by @michaelrabone
@michaelrabone
Top 20 Light Conditions Example: Moonlight Portrait taken only with moonlight… — Midjourney prompt by @michaelrabone
@michaelrabone
Cinematic Street Photography 5 Midjourney Motion Style Mix Anya Taylor-Joy,… — Midjourney prompt by @michaelrabone
@michaelrabone
QT Sci-Fi A lone astronaut standing on a cliff overlooking a vast alien… — Midjourney prompt by @JasonHeatony7
@JasonHeatony7
Important to remember the difference between 'Desert' and 'Dessert' in your… — Midjourney prompt by @michaelrabone
@michaelrabone
Exploring black and white skin together With Midjourney style and my profile… — Midjourney prompt by @michaelrabone
@michaelrabone
A charming frog perched on a broad green lily pad in a beautiful, tranquil… — Midjourney prompt by @heathergreen
@heathergreen
1. Pentax KM camera 2. Pentax-A 50mm F1.2 lens 3. Ilford HP5 Plus 400 35mm film… — Midjourney prompt by @michaelrabone
@michaelrabone
Widescreen 16:9 Cinematic Lights Click on images for full view Anya Taylor-Joy,… — Midjourney prompt by @michaelrabone
@michaelrabone
Midjourney V8.2 : Watercolor portrait illustration. A 20-year-old seductive… — Midjourney prompt by @VIBEQUIRKLABS
@VIBEQUIRKLABS
1. Kodak 35mm film 2. Long exposure (motion) 3. Creative golden hour light… — Midjourney prompt by @michaelrabone
@michaelrabone
Portrait taken only with moonlight of elderly man with long grey beard standing… — Midjourney prompt by @michaelrabone
@michaelrabone
살짝 수위 높은 액션 장면에 도전해 보는 중🩸 kunoichi, fullbody, simple background --ar 3:4… — Midjourney prompt by @imhealmachine
@imhealmachine
character in environment --ar 9:16 --sref 124935953 3650745203 829029026 — Midjourney prompt by @airina_xyz
@airina_xyz
Photography & Vintage Style Mix Portrait of Scarlett Johansson, Kodak 35mm… — Midjourney prompt by @michaelrabone
@michaelrabone

Midjourney prompts by category

Choose a collection based on the image brief, not only a broad style word. A portrait needs a clear subject, crop and light direction; a poster needs hierarchy and negative space; a product image needs the object and its materials to stay legible. These categories link to existing PicGens collections, each backed by its own public examples.

How to write a good Midjourney prompt

Build the prompt in a readable order: subject, action or pose, setting, composition, light, visual treatment and only then parameters. The subject should say what the viewer sees; the action should clarify what is happening; the setting and composition place it in the frame. Add style and lighting cues that reinforce the same direction instead of stacking adjectives that pull the image in different directions.

  • Subject: name the person, object or scene and the detail that should attract attention first.
  • Action and setting: describe a visible moment and the environment that explains it.
  • Composition: say close-up, full-body, overhead, centered, profile or wide establishing shot when framing matters.
  • Lighting and color: use a light source and direction, then a restrained palette or contrast relationship.
  • Finish: choose one medium or visual tradition, such as editorial photography, ink illustration or a paper-cut diorama.

Before and after: make a prompt more specific

These short pairs are illustrative rewrites, not copied gallery prompts. They show how adding one concrete instruction at a time makes a brief easier to interpret. Keep the subject and intent intact while improving framing, light and hierarchy.

Starting pointMore specific direction
A cinematic portraitA close-up profile portrait at a rain-streaked window, cool window light on the face and a warm practical lamp behind.
A product posterA vertical perfume poster with one bottle centered on black glass, a narrow rim light, and a short headline in clear space above.
A fantasy landscapeA wide view of a stone bridge over a misty valley at dawn, bridge leading toward a distant tower, pale sky and dark foreground framing.

Midjourney parameters cheat sheet

Parameters belong at the end of the prompt. Start with the scene description, change one control at a time, and record the version and settings when comparing runs. Current availability can differ by model version, so use the official compatibility notes before copying an older prompt unchanged.

ParameterWhat it changes and version note
--arSets the width-to-height ratio, for example 3:2 landscape or 2:3 portrait. The ratio describes shape, not final pixel dimensions.
--stylize / --sAdjusts how much the model applies its own aesthetic treatment. Lower values follow the wording more literally; higher values allow more interpretation.
--chaos / --cChanges the variety among the initial results. Higher values make the options less alike and less predictable.
--srefApplies a Style Reference image or code to guide visual style. Style References remain supported in V8.2; the code behavior and style-version context can vary.
--crefLegacy Character Reference syntax. V8.2 replaces Character Reference with the Edit Model; use current reference-image workflows for V8.1/V8.2.
--vSelects an image-model version when needed. Midjourney documents V8.2 as the current default; older prompts can explicitly name an earlier version.
--nijiSelects a separate Niji model series for anime and illustrative work. Niji 7 is a separate version, not a V8.2 image-model parameter setting.

Use parameters to test one change at a time

A parameter is most useful when it answers a specific question about the image. If the framing is wrong, change --ar and leave the scene alone. If the image is too interpretive, reduce stylization rather than rewriting every noun. If the first set of options is too similar, try a small change to chaos. This makes each run easier to compare and keeps a prompt from becoming a list of controls that no longer has a clear visual brief.

  • Frame first: choose a portrait, square or landscape ratio that gives the subject room to read.
  • Control interpretation second: adjust stylize only after the subject, pose and setting are explicit.
  • Explore alternatives third: change chaos when you want the initial options to diverge more.
  • Keep style separate from subject: use --sref for the visual treatment and the written prompt for what should appear.
  • Record version and settings with a result so a later comparison uses the same baseline.

How to reuse a style-reference code

The examples below come from distinct public Midjourney prompt rows in this catalogue. They are ordered by the likes captured in the current snapshot, so the list is a way to discover real prompts that include a style-reference code. It is not a ranking of code quality, a test of reproducibility, or a claim that the code alone caused a result. Follow each card to inspect the complete prompt and source.

  • Keep the code with the subject text when you want to study the original combination.
  • To explore a code on another subject, change the subject phrase and keep the rest of the prompt simple.
  • For a style-reference image instead of a numeric code, follow Midjourney’s current Style Reference instructions.

Midjourney vs Nano Banana vs GPT Image

The tools use different prompt workflows, so compare them against the task you need to finish. Midjourney prompts commonly combine descriptive scene text with parameters and Style References. Nano Banana examples often describe a direct edit to an attached image. GPT Image 2 examples can be useful when the brief includes a precise layout or short text. These are workflow distinctions, not a benchmark: results change with model versions, account settings and inputs.

A simple Midjourney prompt review checklist

Before you run a prompt, read it once from the subject outward. The first phrase should tell the model what the image is about. Later phrases should refine the scene instead of introducing competing subjects. Check that the camera position matches the intended crop, that the light source can plausibly illuminate the scene, and that the palette supports rather than obscures the focal point. If text needs to appear in the image, keep the requested wording short and inspect spelling in the output. If the image needs to match an existing reference, use the right current reference workflow and describe the new subject clearly.

  • Remove repeated adjectives that do not add a visible detail.
  • Replace abstract praise such as “beautiful” with a concrete choice about pose, material, color or light.
  • Move the aspect ratio and other parameters to the end of the prompt.
  • After generation, compare the result against the brief and change the single clearest miss first.

Work through a portrait prompt in three passes

Begin with a clear photographic brief: a waist-up portrait of a violin maker in a small workshop, turned toward a window, holding an unfinished instrument. Decide which element should attract attention first and whether the workshop needs to be recognizable. For a first pass, keep the clothing, background and light description simple. Review the pose and the relationship between the hands and instrument before adjusting the visual treatment. If the crop loses the instrument, ask for more room below the shoulders; if tools dominate the scene, reduce the background detail. A second pass can change the light to a soft side source while retaining the pose and framing. A third pass can add a restrained color treatment or surface texture. Keep the text and settings beside each output so the comparison records what actually changed. This is an illustrative writing exercise, not a claim that PicGens generated or tested an accompanying image.

Turn a product idea into a usable poster brief

Separate the product, environment and typography into distinct instructions. For example, describe a simple bottle on a matte stone plinth, a warm background, a narrow side highlight, and clear empty space above for a short headline. The product needs enough room to remain readable after cropping. Keep the title wording separate from the description of the bottle, and say which text should appear if lettering is part of the request. If the output changes the product shape or invents a label, correct that visible miss before increasing the number of decorative elements. Compare the full image and a small thumbnail: a design that looks detailed at full size can lose its hierarchy in a feed. For an actual campaign, check product details and text against approved source material, then finish any inaccurate typography in a design editor. The prompt supplies a visual brief; it does not replace proofing or product review.

Record the context of a style-reference experiment

When studying a shared style code, keep the original prompt, source link and recorded model label together. A code can appear alongside several parameters or reference images, so the visible result should not be attributed to that number alone. Start a comparison by holding the subject and composition steady, then change one style instruction or control. Inspect color, texture, edge treatment and light separately instead of describing the entire result as better or worse. If a style makes the subject hard to read, simplify competing style language before changing the subject. Save the useful result with its actual settings and date. That record makes a later adaptation easier to understand even if the available model version or code interpretation has changed. The catalog examples below are selected from published records that pass the current detail index policy; a source link remains the place to inspect the creator’s context.

Popular --sref Style Codes

Each card below is a public Midjourney example whose shared prompt contains the displayed code. These 10 examples pass the current detail index policy and are ordered by the likes recorded in this catalog snapshot; that ordering is a discovery aid, not a guarantee about the style code or its output.

Frequently asked questions

Are these Midjourney prompts free to use?
You can copy the prompt text shown on PicGens. A public post is not a licence to republish the creator’s prompt or image; check its source and the terms that apply to your use.
How do I copy a Midjourney prompt and use it?
Open a gallery card, copy its shared prompt, then paste it into Midjourney’s Imagine bar or an /imagine prompt. Review the version and parameters before running it.
What does --sref mean in Midjourney?
It adds a Style Reference code or image to guide visual qualities such as color, texture and lighting. The reference does not reproduce the original objects; keep the text prompt focused on the new subject.
What is the best aspect ratio for Midjourney portraits?
Choose a portrait-shaped ratio when you want more vertical framing, then adjust the crop to the final use. Midjourney’s --ar controls proportion rather than a specific pixel size, and the best choice depends on the composition.
Can I use Midjourney prompts in Nano Banana or ChatGPT?
The descriptive subject, scene and lighting language can be adapted. Remove or rewrite Midjourney-specific parameters, and turn reference-image instructions into the format supported by the other model.
Why does my Midjourney result look different from the example?
The original may use a different model version, reference image, personalization, settings or rerun. PicGens preserves the source prompt where available, but does not guarantee identical outputs.
How do I make Midjourney images more realistic?
Describe a photographic setup: subject, lens or shot distance, light source, environment and realistic material details. Change one instruction at a time and inspect the result; no prompt wording guarantees photorealism.