Prompting guide
How to prompt Grok Image
Grok Imagine prompt guide: how to prompt Grok Image with xAI's subject-first order, a check-then-edit loop, and copyable photo prompt examples.
Quick start
Your first great Grok Image prompt
Everything a beginner needs in one minute: what to know first, the rules that matter most, a template to fill in, and prompts to copy.
What to know first
Grok Image is judged first on one thing: is the subject in the picture the subject in your prompt? So open with one sentence that names it, the form it is in, what holds it, and where it is. Then add the details, the light, the colors, and the crop: four to six sentences of plain prose describe one strong photo. Upload an image and the prompt becomes an edit.
The rules that matter most
Open with the subject sentence
"A matte sage-green ceramic pour-over dripper sitting on a glass carafe half full of coffee, on a pale oak counter against white square tile." Opening with mood, film stock, or an avoid list turns the subject into a stand-in.
Name each object exactly, and count it
"A tan leather satchel", not "a bag"; "an enameled cast-iron skillet", not "a dish"; "two espresso cups". Grok draws the object you name.
Make the light bright enough to read
Give the light a source, a direction, and a hardness: "soft window light from the left, bright and warm". Words like "underexposed" darken the subject.
Put the change first in an edit
"Replace the metal watch strap with a brown leather strap. Keep the same wrist, the same angle, and the same desk." An edit that opens with "keep everything" barely moves.
Check, then fix one thing
If the subject is wrong, sharpen the subject sentence and generate again. If only a detail or two is off, upload the result and edit just that.
Fill in this template
Replace each part in square brackets with your own idea, and leave out a part you do not need.
[Subject, its form, what holds it, and where]. [Key details and counts], seen from [camera angle]. [Props and surfaces]. [Light]. [Colors of the subject and of the background]. [Crop and focus], from the same angle.A stack of three buttermilk pancakes on a white enamel plate on a marble café table. Melting butter slides off the top, blueberries spill down the side, and maple syrup runs over the edge, seen from a 45-degree three-quarter view at table height. A steel fork, a linen napkin, and one cup of black coffee sit beside the plate. Bright morning sun from the right with soft fill, glossy highlights on the syrup. Golden pancakes and deep blue berries; a white marble table and a pale green wall. Cropped just inside the plate rim, the top pancake sharp, from the same 45-degree three-quarter view.[Subject, its form, what holds it, and where]One sentence: a stack of three buttermilk pancakes on a white enamel plate on a marble café table.[Key details and counts]Each element by name and form: melting butter, blueberries spilling down the side, maple syrup running over the edge.[camera angle]Where the eye enters: a 45-degree three-quarter view at table height.[Props and surfaces]Two or three named things: a steel fork, a linen napkin, one cup of black coffee.[Light]Source, direction, and hardness: bright morning sun from the right with soft fill.[Colors of the subject and of the background]Two statements: golden pancakes and deep blue berries; a white marble table and a pale green wall.[Crop and focus]What the frame cuts off and what stays sharp: cropped just inside the plate rim, the top pancake sharp.
Prompts to start from
A product shot
A pair of white leather low-top sneakers with gum rubber soles, shot from a low three-quarter angle. The laces are loosely tied, the toe boxes show soft creases, and the left shoe sits slightly ahead of the right. They stand on pale gray concrete against a warm off-white wall. Hard morning sun from the right casts long, sharp shadows, with soft fill on the shaded side. 50mm product photograph from a low three-quarter angle, both shoes sharp, cropped just above the heel tabs.Why it works: The subject comes first in its exact form, the angle is stated twice in the same words, and the light has a source, a direction, and a hardness. A 3:2 aspect ratio suits it.
A portrait at work
A woman in her sixties with short silver hair and reading glasses pushed up on her head, laughing while she pots a small olive tree at a wooden workbench. Her hands are dusted with dark soil and her sleeves are rolled to the elbow. Terracotta pots and a steel trowel sit on the bench beside her. Soft window light from the left, bright and warm, with gentle shadows. 35mm photograph at eye level, framed from the waist up, her face in sharp focus.Why it works: Visible facts instead of adjectives (the glasses, the soil, the rolled sleeves) and two named props make a person you could point at. Try a 2:3 aspect ratio.
Food that fills the frame
A vegetarian Greek salad in a warm cream ceramic bowl, shot from a slight overhead angle. Cucumber, tomato, thick red onion, irregular feta crumbles, and wrinkled black olives overlap in olive oil, with oregano on the tomatoes. The bowl sits on butter-yellow linen over terracotta; a warm pita is partly visible at the right edge and a fork rests on the rim. Hard summer sun and on-camera flash, bright and vivid, with warm shadows. 35mm photograph from a slight overhead angle, the food filling the bowl, one frame.Why it works: Every ingredient is named in the form it takes, and the bright, hard light keeps the subject readable. Use 4:3.
An edit with the change first
Replace the metal watch strap with a brown leather strap with cream stitching. Keep the same wrist, the same shirt cuff, the same angle, and the same desk.Why it works: Upload the photo and set the aspect ratio to match it. The change leads, and one closing sentence keeps everything else in place.
A change of style
Render this as a pencil sketch with detailed cross-hatched shading on off-white paper. Keep the same people, poses, and framing.Why it works: Upload the photo. A style change describes the look you want, then keeps the subject and the layout.
Advanced
Advanced Grok Image prompting
Techniques for more control once the basics work.
Which Grok Image tier
Grok Image comes in three tiers that read the same prompt. Grok Image is the quick one, for drafts and volume. Grok Image Quality is the default. Grok Image 2.0 is the newest, and the prompting advice xAI publishes, which this guide follows, was written for it.
On Grok Image 2.0, the Quality setting chooses Low or Medium. Use Medium at 2K when fine detail has to be recognizable: a watch face, a fabric weave, a product label.
- xAI is retiring the Quality tier on November 2, 2026. After that, Grok Image Quality is drawn by Grok Image 2.0 at Low quality, so start a new series on Grok Image 2.0 to keep its look steady.
- Sketch the composition on Grok Image, then run the final on Grok Image 2.0 with the same words.
The six-part order
xAI writes every prompt in the same six parts, in the same order, and a series stays consistent when every prompt keeps that order. The quick-start template follows it.
Write prose that describes one photograph. A new image has no "before", so "make the coat darker" means nothing in it; write "a charcoal wool coat".
- Subject: what it is, the form it is in, what holds or supports it, and the setting, in one opening sentence.
- Details and focal point: every key element by name and in the form it takes ("a wool coat in heavy twill with a visible weave", "coffee in a glass carafe, dark amber, lighter at the edges"), how many of each, and the camera angle.
- Art direction: the color fields and two or three named props ("oak counter, white tile, linen towel, one brass spoon").
- Light: source, direction, and hardness; what the highlights and shadows look like; an exposure bright enough to read the subject.
- Color: the subject's colors and the background's colors, as separate statements.
- Composition: the crop, what the frame cuts off, and what stays sharp. One scene.
- Give the camera angle with the details, then repeat it in the same words in the composition.
Generate, check, then edit
Check every result against your prompt, point by point: the right subject, in the form you described; every named element visible; every count right; each object the one you named (a plate, not a bowl; a tote, not a backpack); and bright enough to read. A perfect background with the wrong subject is a miss.
Then choose the fix. When the subject itself is wrong (another object, a missing element, a wrong count), sharpen the subject sentence and generate again. When the subject is right and only one to three details are off, upload the result and edit those details. Each pass changes only what failed; keep the rest of the prompt word for word.
- When a fix does not land, change the words and point at the part: "the leather shows creases where the strap bends", not "more realistic" again.
- Keep each prompt next to its image, so the next pass starts from the exact words.
Edits that move the right amount
Lead an edit with the changes, then add one sentence that keeps the same subject, angle, crop, and prop placement. An edit that opens with "keep everything" barely moves. An edit that pastes in a whole new prompt restages the scene: the subject drifts, props double, and objects change shape.
Chain edits by uploading each result for the next one, and set the aspect ratio to the uploaded image's shape: it starts at square.
- For a style change, describe the target look in the edit: "Render this as a pencil sketch with detailed cross-hatched shading."
- For an ad series, edit one product photo into several scenes at the shape you will post, asking each time to keep the product exactly as it is.
One product, a new scene
Place this bottle of olive oil on a sunlit stone windowsill in a Mediterranean kitchen, with a bowl of lemons and a sprig of rosemary beside it. Keep the bottle, its label, its cork, and its proportions exactly as they are.Why it works: Upload the product photo. The new scene comes first and one closing sentence locks the product, so every scene in the series shows the same bottle.
References and Characters
Grok Image takes a few images at once. Mention saved Characters and References by their @handle and uploads as @image1, @image2. Mage numbers them in one list (your uploads first, then Characters, then References) and swaps each mention for its number, so with one upload, @ana becomes "character from image 2". Say what each image gives: "Put the sneaker from image 2 on the concrete step in image 1."
To keep a character the same across scenes, save them as a Character and mention the @handle in every prompt. Or use xAI's storyboard method: generate the first scene, then make each other scene as an edit of it that uses it only as a reference for the subject and the style, and asks for the new scene and camera angle.
The next scene of a storyboard
Use image 1 only as a reference for the woman and the illustration style. New scene: she stands on a windy cliff path above the sea at dawn, her red scarf blowing to the left, seen from a low angle behind her right shoulder. Keep her face, her hair, the red scarf, and the ink-and-wash style the same.Why it works: Upload the first scene. Saying what the image is for, and only that, stops Grok from copying its layout, while the new scene and angle come from the words.
Light, materials, and texture
Describe light by its source, direction, and hardness, and objects by their materials. Grok paints what the words name, so name the physical thing: "brushed steel with fine horizontal grain", not "not shiny".
- State each texture once, on the surface it belongs to. Texture words repeated across a prompt get painted onto the wrong things.
- Leave defect words out: "chromatic fringing", "dust", and "scratches" are drawn as defects.
- Styles beyond photography work too: anime, oil painting, pencil sketch. Name one.
| Write | What it does |
|---|---|
45-degree three-quarter view at counter height | The product angle: the top and the front both visible. |
straight-down overhead | A flat lay, seen from directly above. |
low three-quarter angle | Slightly below and to the side, so the subject stands tall. |
eye level, framed from the waist up | A natural portrait crop. |
hard afternoon sun from the left with on-camera fill flash | Crisp shadows, with the dark side still readable. |
soft window light from the left, bright and warm | Gentle, directional light for portraits and food. |
bright and vivid, hard sun, flash, warm shadows with texture | An exposure bright enough to read every detail. |
Text in the image
Keep text to a short label or sign: quote the exact words and say where they sit and in what lettering. xAI publishes no advice on typography, so for posters, menus, or anything led by its words, use GPT Image 2.5 Flare.
Common mistakes
Avoid this, do this instead
Avoid: Opening with mood, film stock, or "avoid…".
Moody, cinematic, Kodak film look, no clutter: a coffee setup on a counter.Instead: The subject sentence first.
A matte sage-green ceramic pour-over dripper sitting on a glass carafe half full of coffee, on a pale oak counter against white square tile.Avoid: "Slightly underexposed" or "crushed shadows".
Instead: Bright, readable light with texture in the shadows.
Avoid: Defect words: "chromatic fringing", "dust", "scratches".
Instead: Nothing: they get drawn.
Avoid: A long avoid list.
Instead: A count or a specific object, paired with what you want: "one stainless spoon resting on the saucer".
Avoid: "Make the coat darker" in a new prompt.
Instead: "A charcoal wool coat": a new image has no before and after.
Avoid: "A bag", "a dish".
Instead: "A tan leather satchel", "an enameled cast-iron skillet".
Avoid: Repeating "more realistic".
Instead: The exact part that looks wrong, and how it should look.
FAQ
Grok Image prompting questions
How long should a Grok Image prompt be?
Does Grok Image support negative prompts?
How do I keep a character consistent in Grok Image?
Can Grok Image do text?
Which Grok Image tier should I use?
Why does my Grok edit barely change anything?
The background is right but the subject is wrong. Edit or regenerate?
Prompting guide
More prompting guides
How to prompt Anima
How to prompt Anima on Mage: Danbooru tag order, rating tags, strong weights, escaped parentheses, and Anima prompt examples for characters and scenes.
How to prompt Chroma
Chroma prompt guide: how to prompt Chroma with one caption-style paragraph, use its negative prompt, tune steps and guidance, and copy Chroma prompt examples.
How to prompt Flux 2
Flux 2 prompt guide: how to prompt Flux 2 Dev with subject-first word order, JSON prompts, exact hex colors, and multi-image edits, with examples.
How to prompt GPT Image 2
GPT Image 2 prompt guide: how to prompt GPT Image 2.5 Flare and Sunburst for exact text, precise edits, and consistent characters, with examples.
How to prompt Guava
How to prompt Guava, Mage's photorealism model: a quick start, a fill-in template, and expert tips for portraits, skin, light, lenses, edits, and text.
How to prompt HiDream
HiDream prompt guide: how to prompt HiDream with plain sentences, object attributes and relations, and quoted sign text, plus HiDream prompt examples.
Make your first image with Grok Imagine
New accounts get 300 free Gems, enough to try Grok Image Quality. No install, no credit card.
Try Grok Image Quality