The /imagine trick that started a thousand cheat sheets
The secret nobody talks about: the most powerful image prompts aren't paragraphs. They're not even single sentences.
They're slash commands — shorthand modifiers you chain together like building blocks.
/portrait /1980s /bokeh /ar-3-4
That's a complete, professional image prompt in five seconds. And it works across Midjourney, DALL-E 3, Leonardo.ai, SDXL, Flux, and ChatGPT.
Here's how this shorthand revolution started — and why you should care.
The birth of /imagine
In March 2022, Midjourney opened its Beta to the public. Everyone got the same first instruction: type /imagine.
Not a long prompt. Not a style guide. Just one slash and a word.
Within weeks, Discord servers were exploding with people sharing their /imagine prompts — and a pattern emerged. The most prolific generators weren't writing paragraphs. They were dropping a few keywords, a style, a camera modifier. Something like:
/imagine prompt: a woman made of smoke / cyberpunk / neon / cinematic lighting / ar 16:9
That's not a sentence. It's a command sequence. And it was faster than English.
Fast-forward to 2026. The /slash convention has gone multi-platform:
-
ChatGPT image gen uses
/imagine(inherited from Midjourney muscle memory) -
Leonardo.ai supports
/model,/style,/arshorthand -
Stable Diffusion web UIs use
/ar,/q,/sameseed - DALL-E doesn't natively use slashes but follows the same compositional pattern
Each platform has 50–200 native modifiers. Combined? 500–1000+ composable commands.
Why slash commands beat full prompts (most of the time)
There's a trade-off every image generator runs into:
| Full Prompt | Slash Command |
|---|---|
| ✍️ Write a paragraph | ✅ Drop a code |
| 📱 Copy-paste is fragile | 💾 Save and reuse |
| 🔄 Hard to edit | 🔗 Combine freely |
| ❓ 800 tokens | 🔢 5 tokens |
Slash commands are deterministic shortcuts. You learn them like muscle memory — type /cyberpunk /neon /ar-16-9 and you know what you'll get before the image even renders.
The sweet spot: build a library of 20–30 core commands you know by heart, then chain them per project. That's what this 1027-command handbook distills.
The 5 categories you'll use 80% of the time
Not all 1027 commands are equal. Based on analyzing which ones ship most often:
1. 📐 Foundation Modifiers (30 commands)
These control the basics — the "what format" question.
/ar-16-9 → 16:9 (YouTube thumbnail)
/ar-1-1 → 1:1 (Instagram square)
/quality-high → 2x GPU / max detail
/stylize-medium → balanced artistic flair
/no-text → bans text in output
/seed-pin → reproducible seed
2. 🔦 Perspective / Camera (30 commands)
The "from where" question.
/lowangle → low camera angle
/overhead → top-down
/birdseye → looking straight down
/pov → first-person point of view
/ultra-wide → fish-eye wide angle
3. 🎨 Style & Medium (30 commands)
The "how does it look" question.
/photoreal → photo realism
/oilpainting → oil on canvas
/cyberpunk → neon-drenched sci-fi
/lowpoly → geometric faceted look
/glitch-art → digital corruption aesthetic
4. 💡 Lighting (26 commands)
The "mood" question.
/goldenhour → warm evening light
/neonlight → neon tubes/glow
/volumetric → visible light rays
/ring-light → creator/streamer setup
/lowkey → high contrast shadows
5. 📱 Output Format (20 commands)
Where it lives — the "what shape" question.
/youtube-thumbnail → 1280×720 ready
/tiktok-cover → 9:16 mobile
/poster → 3:4 tall poster
/book-cover → 2:3 A4-ish
/banner → 16:7 wide
Master these 136 commands, and you can compose ~8,000 valid combinations without ever writing a full-sentence prompt.
Advanced: chaining for complex results
Here's where slash commands get dangerous. You're not limited to one per category:
/imagine: a lone astronaut / cyberpunk / neon / lowangle / goldenhour / rain / motionblur / ar-9-16 / no-text
That's one command chain producing a cinematic, specific, ready-to-post image. The astronaut is a modifier — it slots into any scene chain.
Power users chain 3 foundation + 3 style + 2 perspective + 1 lighting + output = 9 commands. That's more creative precision than a 100-word paragraph, and infinitely more editable.
What's in the full handbook
I didn't stop at the 136 power commands. The complete handbook covers all 1027 commands across 28 categories:
🎯 Foundation (aspect, quality, model, seed)
📐 Technical (CAD, blueprint, exploded view, cutaway)
🧪 Science/Medical (DNA, anatomy, microscope, MRI)
📊 Infographics (charts, flowcharts, dashboards)
📦 Product (product shots, packshots, unboxing)
📣 Advertising (ads for Meta, Google, billboards)
📱 Social Media (IG, TikTok, YouTube, LinkedIn)
📸 Photography (film stocks, lenses, bokeh)
🎬 Cinematic (movie posters, noir, steampunk)
🌅 Lighting (golden hour, neon, rim light)
🌍 Camera Angles (drone, bird's eye, Dutch tilt)
🎨 Art Styles (oil, manga, pixel art, stained glass)
🧊 3D/CGI (Blender, Octane, ray tracing)
🏢 Architecture (floor plans, modern, brutalist)
👤 People/Fashion (headshots, runway, vintage)
🏎️ Vehicles/Machines (cars, rockets, aircraft)
🧠 Concept/Creative (cyberpunk, dystopian, surreal)
🏷️ Branding/Design (logos, posters, style guides)
🖼️ Image Editing (restore, inpaint, sky replace)
🗺️ Maps/Travel (topographic, fantasy, metro maps)
🕰️ Historical/Vintage (decade styles, daguerreotype)
🎨 Colors & Palettes (duotone, neon, earth tones)
🧱 Textures & Materials (concrete, marble, holographic)
🌿 Nature/Elements (forest, storm, aurora)
🦁 Animals/Creatures (wildlife, dragons, mythical)
🧸 Objects & Everyday (typewriters, books, vinyl)
💭 Emotions/Concepts (nostalgia, tension, wonder)
🌀 Abstract/Conceptual (fractals, mandalas, neural nets)
It also includes:
- ✅ How-to-use cheat sheet (page 3) — 4 example combos with full breakdowns
- ✅ Compatibility matrix — which commands work on Midjourney vs DALL-E 3 vs Leonardo vs SDXL vs Flux vs ChatGPT
- ✅ 10 quick workflows — "turn your photo into anime," "generate a product catalog," "create a YouTube thumbnail set"
Grab the complete handbook
Building this library was a rabbit hole — I dug through Midjourney's v7 changelog, DALL-E 3's latest parameter updates, Leonardo.ai's 2025 feature drops, and the SDXL/Flux ecosystem to make sure every one of these 1027 commands has an accurate compatibility tag.
The result is a 34-page PDF — not a wall of text, but a scannable command matrix. Print it. Bookmark it. Stick it beside your monitor. In a week, /cyberpunk /neon /ar-9-16 /no-text will feel as natural as writing your name.
→ Get the AI Slash Commands Handbook on Gumroad
It's the one PDF you'll reach for every time you fire up an image generator — because once you start chaining commands, full-sentence prompting starts to feel archaic.
This handbook covers slash-command shorthand used by Midjourney, DALL-E 3, Leonardo.ai, Stable Diffusion, Flux, and ChatGPT image generation. Platform names and versions referenced are accurate as of August 2026.
Bonus: 5 chaining recipes for any style (dev.to edition)
Copy-paste any of these into Midjourney, Leonardo, or ChatGPT /imagine:
1. Cyberpunk portrait
/portrait /cyberpunk /neon /rain /ar-9-16 /stylize-high /quality-high
2. Product hero shot
/producthero /studio /softbox /glass-render /ar-16-9 /no-text /quality-high
3. Movie poster
/movieposter /cinematic /dramaticlight /ar-2-3 /stylize-medium /filmgrain
4. Educational diagram
/scientificdiagram /infographic /3dmap /colorize /ar-16-9 /labelled
5. Social media thumbnail
/youtube-thumbnail /viralthumbnail /boldad /highcontrast /ar-16-9 /no-text
Each chain = 1 foundation + 1 output format + 2–3 style/perspective modifiers. The full handbook gives you 1027 commands to remix however you want.
Top comments (0)