I’m Testing 200+ AI Image Shortcodes — One Photo, Every Command
I found hundreds of viral AI image commands. Instead of just saving them, I want to test what every one actually does to a photograph.
I've been seeing more and more posts claiming that you can use short commands such as:
/cinematic
/motionblur
/topview
/goldenhour
/anime
/bokeh
and many others to transform photographs with AI.
The idea is simple:
Upload a photo → use a shortcut → let the AI transform it.
I came across a large collection from G Prompt that describes itself as “200+ Viral Shortcodes for AI Shortcut Prompt Editing.” The page presents these as image-to-image shortcuts for ChatGPT, covering everything from photo repair and background changes to camera angles, lighting, effects, transformations, and social-media-ready edits.
I also found another collection of 99 ChatGPT AI image commands covering camera perspectives, lighting, weather, photography styles, artistic styles, environments, materials, and visual effects.
That gave me an idea.
Instead of collecting these commands, I’m going to test them.
All of them.
1. What exactly am I trying to find out?
The internet is full of lists called things like:
“Best AI image commands”
“Viral ChatGPT image prompts”
“Secret AI photo-editing codes”
“100+ AI image shortcuts”
But seeing a command on an Instagram carousel is not the same as proving that it works.
For example:
/motionblur
sounds useful.
But what actually happens when I use it?
Does the AI add convincing motion?
Does it preserve the person correctly?
Does it change the face?
Does it alter the background?
Does it produce something I would actually use?
That is what I want to discover.
2. This is going to be a real photo experiment
Instead of generating random images for every command, I want to establish a repeatable experiment.
I'll start with a set of photographs.
Then I'll apply the commands to those photographs.
The objective is to observe:
What changes?
What stays the same?
What breaks?
Which commands are actually useful?
3. My basic testing method
For every command, I'll follow roughly the same process.
Step 1 — Start with an original photograph
The original image becomes the baseline.
Step 2 — Apply one command
For example:
/MOTIONBLUR
Step 3 — Inspect the result
I'll compare the original and edited image.
Step 4 — Record what changed
I'll look at things such as:
subject identity
face
clothing
pose
background
lighting
perspective
colours
realism
unintended changes
Step 5 — Give it a rating
I'll rate the command based on how well it performs its intended job.
4. The most important rule
I don't want to judge an image simply because:
“It looks cool.”
A visually impressive result can still be a bad edit.
For example, suppose /TOPVIEW creates a beautiful image but completely changes the person's body and environment.
That might look interesting.
But it doesn't mean the command worked properly.
So I'll separate:
Visual quality
Does it look good?
from:
Transformation accuracy
Did it actually do what the command was supposed to do?
That distinction is important throughout the experiment.
5. My rating system
I'll use a simple 1–5 system.
⭐ 1/5 — Didn't work
The intended effect wasn't achieved or the image became unusable.
⭐⭐ 2/5 — Weak
Something changed, but the result isn't particularly useful.
⭐⭐⭐ 3/5 — Works
The effect is visible and reasonably useful.
⭐⭐⭐⭐ 4/5 — Very good
The command performs its intended transformation well.
⭐⭐⭐⭐⭐ 5/5 — Excellent
The result is strong enough that I would genuinely use the command again.
I'll also record a separate “Keep” verdict:
✅ Keep
š¤ Situational
❌ Skip
6. I'm going to test commands by category
Rather than testing them randomly, I'll group them.
That should make patterns easier to discover.
CATEGORY 1 — Photo Fixing & Recovery
These commands are intended to improve damaged, weak, blurry, or poorly captured images.
Commands
/PHOTOFIX
/PERSONREMOVE
/OBJECTREMOVE
/BLURFIX
/OUTFITFIX
/HAIRFIX
/LIGHTFIX
/INSTAFIX
/BESTLOOK
/GLOWUP
/MAKEITBETTER
/PHOTORESCUE
/KEEPITREAL
/PROSHOT
/PERFECTSHOT
/SELFIEFIX
/CAMERAREADY
What I'll test
I'll use imperfect photographs and see whether the commands can improve them without destroying the original identity or scene.
CATEGORY 2 — Face & Appearance
Commands
/FACEFIX
/FACEGLOW
/SKINCLEAN
/SKINREAL
/EYEPOP
/HAIRFIX
/HAIRCHANGE
/HAIRCOLOR
/SMILEFIX
/MAKEOVER
/BEARD
/GLASSES
Questions
Does the face remain recognizable?
Does the result stay natural?
Does the AI make changes I didn't ask for?
Does “improvement” actually mean improvement?
CATEGORY 3 — Outfit & Fashion
Commands
/OUTFIT
/OUTFITMATCH
/STYLEUP
/OLDMONEY
/QUIETLUXURY
/CASUALSTYLE
/FORMALSTYLE
/DATELOOK
/PARTYLOOK
/WEDDINGLOOK
Test idea
Take the same photograph and systematically change the clothing style.
Original:
casual
↓
/FORMALSTYLE
↓
/OLDMONEY
↓
/QUIETLUXURY
↓
/WEDDINGLOOK
Then compare the results.
CATEGORY 4 — Removing Unwanted Objects & People
Commands
/PEOPLECLEAN
/CROWDREMOVE
/OBJECTREMOVE
/DISTRACTIONREMOVE
/CLEANBACKGROUND
/BGREMOVE
/MESSYCLEAN
Test
I'll deliberately choose photographs containing distracting elements.
Then I'll see whether the AI can remove them while keeping the surrounding environment believable.
CATEGORY 5 — Background Replacement
Commands
/NEWBG
/BGPRO
/BGNATURE
/BGCITY
/BGSTUDIO
/BGCINEMATIC
/BGLUXURY
/BGBLUR
Test
One photograph.
Multiple environments.
For example:
Original
↓
/BGNATURE
↓
/BGCITY
↓
/BGSTUDIO
↓
/BGCINEMATIC
↓
/BGLUXURY
Then I'll compare how convincingly the person fits into each environment.
CATEGORY 6 — Quality & Lighting
Commands
/HDREAL
/BLURFIX
/DETAILUP
/LIGHTFIX
/COLORFIX
/NIGHTFIX
/DARKTOBRIGHT
/SHARPREAL
/GOLDENHOUR
/SOFTLIGHT
/WINDOWLIGHT
Questions
Does the AI really recover lost detail?
Does /GOLDENHOUR produce believable lighting?
Does /WINDOWLIGHT look physically plausible?
Does increasing “quality” actually improve the image or simply make it look more artificial?
CATEGORY 7 — Camera Angles & Perspective
Commands
/DRONEVIEW
/TOPVIEW
/LOWANGLE
/SIDEVIEW
/BACKVIEW
/WIDEANGLE
/MACRO
/CLOSEUP
/POVSHOT
/OVERTHESHOULDER
/BIRDEYEVIEW
And from the additional image-command collection:
/aerialview
/highangle
/overtheshoulder
/wormseyeview
/extremecloseup
/fisheye
This could be one of the most interesting categories.
A camera angle isn't just a filter.
It can require the AI to reconstruct:
the person's body
the environment
perspective
lighting
depth
objects outside the original frame
So I'll test how much of the image gets reconstructed correctly.
CATEGORY 8 — Photography & Visual Effects
Commands
/BOKEH
/MOTIONBLUR
/LONGEXPOSURE
/FILMGRAIN
/POLAROID
/DOUBLEEXPOSURE
/SILHOUETTE
/REFLECTION
/LENSFLARE
From the additional 99-command collection:
/cinematic
/rain
/storm
/snow
/sunset
/sunrise
/goldenhour
/bluehour
/moonlight
/neonlights
/cyberpunk
/darkmoody
/softlighting
/dramaticlighting
/rimlight
/backlight
/volumetriclight
/godrays
/studio
/filmgrain
/anamorphic
/shallowdepth
/longexposure
/freezeaction
/vintage
/retro
/oldmoney
/luxury
/minimalist
/editorial
/fashion
/streetstyle
CATEGORY 9 — Nature & Environment
Commands
/FOGYFOREST
/MISTYMOUNTAIN
/RAINYFOREST
/SNOWYFOREST
/SUNSETBEACH
/MOONLIGHT
/NATUREPORTRAIT
/UNDERWATER
/SPACE
/JUNGLE
/DESERT
Additional environment commands:
/underwater
/space
/desert
/forest
/mountains
/beach
/cityscape
/nightcity
/futuristic
/postapocalyptic
/fantasy
/scifi
/medieval
/samurai
/noir
/detective
/superhero
CATEGORY 10 — Social Media Ready Images
Commands
/INSTAREADY
/DPREADY
/STORYREADY
/POSTREADY
/REELCOVER
/AESTHETIC
/MIRRORSELFIE
/CANDIDLOOK
/STREETSNAP
Question
Do these commands actually produce a better image for that platform, or do they simply change the visual style?
I'll test whether they affect:
composition
cropping
subject placement
portrait framing
visual hierarchy
CATEGORY 11 — Birthday & Celebration
Commands
/BIRTHDAYWISH
/BIRTHDAYPOSTER
/BIRTHDAYCARD
/BIRTHDAYSTORY
/BIRTHDAYPARTY
/BIRTHDAYQUEEN
/BIRTHDAYKING
/SURPRISEPARTY
CATEGORY 12 — Fun Transformations
Commands
/MINIME
/GIANTME
/CLONEME
/TOY
/ACTIONFIGURE
/LEGO
/MINIATURE
/CLAY
/STATUE
Additional visual-style commands:
/anime
/comicbook
/oilpainting
/watercolor
/sketch
/claystyle
/3drender
/isometric
/miniature
/tiltshift
/papercraft
/glassart
/chrome
/holographic
/doubleexposure
/silhouette
/reflection
/mirror
/fire
/smoke
/glitch
/dreamcore
CATEGORY 13 — Viral Curiosity & Alternate Versions
Commands
/FUTUREME
/CHILDHOODME
/THENNOW
/MYERA
/ALTLIFE
/PARALLELME
/TIMESPLIT
/PORTALME
/DREAMLIFE
/MAINCHARACTER
/PAPARAZZI
/REDCARPET
/BILLBOARDME
/MOVIEVERSION
/DIGITALTWIN
These are especially interesting because they go beyond editing.
They're asking AI to reimagine the subject.
That creates a different question:
How much can I change before the output stops being a transformation of my photograph and becomes a newly generated interpretation?
CATEGORY 14 — Additional Image-Generation Commands
The second command collection I found contains a large set of shorter visual controls that are less about “editing” and more about telling the model what kind of image to create.
Image quality
/4k
/8k
/hdreal
/ultrarealistic
/hreal
Camera perspective
/droneview
/aerialview
/topdown
/lowangle
/highangle
/sideview
/closeup
/extremecloseup
/wideangle
/fisheye
/pov
/overtheshoulder
/birdseyeview
/wormseyeview
Atmosphere
/fog
/mist
/rainynight
/rain
/storm
/snow
Lighting
/sunset
/sunrise
/goldenhour
/bluehour
/moonlight
/neonlights
/cyberpunk
/darkmoody
/softlighting
/dramaticlighting
/rimlight
/backlight
/volumetriclight
/godrays
/studio
/bokeh
/shallowdepth
Motion & photography
/motionblur
/longexposure
/freezeaction
/filmgain
/anamorphic
/vintage
/retro
/polaroid
/disposablecamera
/35mmfilm
/instantphoto
Style
/oldmoney
/luxury
/minimalist
/editorial
/fashion
/streetstyle
/magazinecover
/anime
/comicbook
/oilpainting
/watercolor
/sketch
/claystyle
/3drender
/isometric
/miniature
/tiltshift
/papercraft
/glassart
/chrome
/holographic
/doubleexposure
/silhouette
/reflection
/mirror
/fire
/smoke
/glitch
/dreamcore
15. There is a problem I want to test
Some commands appear to overlap.
For example:
/TOPVIEW
and
/topdown
and
/BIRDEYEVIEW
and
/aerialview
all sound related.
Similarly:
/GLOWUP
/MAKEITBETTER
/BESTLOOK
/PERFECTSHOT
could all produce some kind of general enhancement.
And:
/vintage
/retro
/polaroid
/35mmfilm
may overlap in their visual effect.
That means this experiment isn't only about:
Does the command work?
It's also about:
Are different commands meaningfully different?
16. The same photo will be extremely important
One of my main experiments will be to take the same photograph and run multiple related commands on it.
For example:
Original
One portrait.
↓
Camera tests
/TOPVIEW
/LOWANGLE
/SIDEVIEW
/CLOSEUP
/WIDEANGLE
↓
Lighting tests
/GOLDENHOUR
/BLUEHOUR
/MOONLIGHT
/SOFTLIGHT
/DRAMATICLIGHTING
↓
Style tests
/CINEMATIC
/VINTAGE
/EDITORIAL
/ANIME
/OILPAINTING
↓
Effects
/BOKEH
/MOTIONBLUR
/LONGEXPOSURE
/FILMGRAIN
This creates something much closer to a controlled experiment.
17. I also want to test combinations
After testing individual commands, I'll start combining them.
For example:
Combination A
/cinematic + /goldenhour
Combination B
/cinematic + /goldenhour + /bokeh
Combination C
/cinematic + /goldenhour + /motionblur
Combination D
/editorial + /studio + /shallowdepth
Combination E
/vintage + /35mmfilm + /filmgain
The question is:
Does stacking commands give me more control?
Or:
Does it simply make the result unpredictable?
18. I want to find the limits
I'll deliberately push the commands.
For example:
Three commands
Usually manageable.
Five commands
More constraints.
Ten commands
Potentially conflicting instructions.
At some point, I want to know whether the AI starts ignoring some of the commands.
That could reveal something important about how these shortcuts actually work.
19. Original prompt vs shortcut
I also want a control test.
For example:
Shortcut
/goldenhour
versus:
Normal instruction
“Change the lighting to warm golden-hour sunlight while preserving the subject, clothing, facial identity, environment, and realistic shadows.”
If the results are basically the same, the shortcut is valuable mainly as a convenience.
If the result changes significantly, that's more interesting.
20. My experiment scoreboard
I'll maintain a table like this:
| # | Command | Category | Works? | Image Quality | Control | Rating | Verdict |
|---|---|---|---|---|---|---|---|
| 001 | /MOTIONBLUR | Effects | — | — | — | — | — |
| 002 | /TOPVIEW | Camera | — | — | — | — | — |
| 003 | /GOLDENHOUR | Lighting | — | — | — | — | — |
| 004 | /BOKEH | Effects | — | — | — | — | — |
| 005 | /OUTFIT | Fashion | — | — | — | — | — |
Eventually, every command gets a result.
21. What I will record for every command
For each experiment, I want to save:
Original image
The starting photograph.
Command
The shortcut being tested.
Input
What I asked the AI to do.
Output
The generated image.
What changed
A short observation.
What went wrong
Any unwanted changes.
Rating
1–5.
Verdict
Keep / Situational / Skip
This will make the final project much more useful.
22. The categories will help me discover patterns
By the end, I should be able to answer things like:
Best category
Which category produces the most reliable results?
Most reliable command
Which command almost always does what it claims?
Biggest failure
Which command repeatedly produces poor results?
Biggest surprise
Which command turns out to be much better than expected?
Most redundant commands
Which commands are basically doing the same thing?
Best combinations
Which shortcuts work especially well together?
Most destructive commands
Which commands change the original image too much?
23. The final goal
I don't want to finish with:
“I tested 200+ commands.”
I want to finish with something much more useful.
Maybe I'll end up with:
200+ commands
↓
150-ish actually visible unique commands
↓
100 useful
↓
50 really good
↓
20 commands I would actually reuse
↓
10 combinations that work exceptionally well
The exact result will depend on the testing.
24. What could come from this?
This experiment can eventually become:
A public case study
A complete record of testing viral AI image shortcuts.
A visual library
Original image → edited image.
A personal shortcut library
The commands that survive the experiment.
A comparison database
Same photo, multiple commands.
A prompt-testing framework
A repeatable method for testing future image commands.
Future content
The experiment can later be turned into articles, image comparisons, carousels, videos, or other content.
But the main purpose is simple:
I want to know what actually works.
25. Credit & sources
The main photo-editing command list I'm using for this experiment comes from G Prompt's “200+ Viral Shortcodes for AI Shortcut Prompt Editing,” published August 25, 2026. The page describes the collection as photo-editing codes for ChatGPT image-to-image workflows.
Original source:
200+ Viral Shortcodes for AI Shortcut Prompt Editing — G Prompt
I am also using the additional 99 AI image commands shown in the visual reference I collected, covering camera angles, lighting, atmosphere, photography styles, environments, art styles, materials, and visual effects.
The purpose of my project is not simply to reproduce these lists.
The purpose is to test them on real photographs, compare the results, document failures and successes, and build my own evidence-based collection of useful commands.
26. Where I'm starting
I'm starting with a photograph.
Not a perfect photograph.
Not a specially prepared AI image.
A normal photograph.
Then:
One command.
One result.
One comparison.
Then another.
And another.
Until I have tested the entire command collection.
The experiment
Take the viral AI image commands people are sharing online. Test every one on photographs. Compare what actually happens. Find the useful commands, the useless commands, the duplicates, the surprising ones, and the combinations that really work.
I don't know yet whether these shortcuts are genuinely powerful.
I don't want to assume they are.
I'm going to test them and find out.
Comments
Post a Comment