An AI prompt keywords cheat sheet, organized by what each keyword does

The core idea in one paragraph

A prompt keywords cheat sheet is only useful if it tells you what each word does, not just that it exists. Most cheat sheets are flat lists you copy without understanding, and copying words you do not understand is how you get the generic AI look everyone complains about. This one is organized by the axis each keyword controls, so you reach for a word because you know which dial it turns. Light keywords turn the mood dial. Camera keywords turn the spatial dial. Style keywords turn the treatment dial. Quality words mostly do nothing on modern models and should usually be cut. Once you understand the axis behind the word, the list becomes a toolkit instead of a spell book.

Why most keyword lists fail you

Almost every popular prompt keyword list has the same flaw. It groups words by surface similarity, like a list of adjectives, without explaining that different words control completely different aspects of the image. The result is that people treat all keywords as interchangeable flavor, stack ten of them, and get an averaged mush. The model only has so much attention to spend, and ten unrelated keywords each claim a slice of it, leaving none dominant enough to actually shape the image.

The fix is to understand that a prompt has a small number of independent axes, and that each keyword belongs to exactly one axis. Within an axis, you pick at most one or two strong choices. Across axes, you combine deliberately. That structure is what separates a prompt that lands from a prompt that wanders.

The axis framework

Before the word lists, here is the framework that makes them useful. Every useful keyword belongs to one of five axes. Knowing the axis is what stops you from stacking contradictory choices.

The light axis

Light keywords are the highest value words in the whole framework. They set the emotional temperature of the image in a single phrase, because each one names a specific lighting setup with a defined mood. We cover the full theory in the lighting article. Here is the operational list.

• Golden hour warm low sun, long shadows, flattering. The safe beautiful default.
• Blue hour cool twilight, no hard shadows, calm and still.
• Overcast soft even light, muted color, clean for products and skin.
• Rembrandt lighting side light with a triangle of light on the shadowed cheek. Moody and sculptural.
• Split lighting half the face lit, half in shadow. Dramatic and confrontational.
• Butterfly lighting light from above front, small nose shadow. Glamorous.
• Low key mostly dark with selective pools of light. Intimate, mysterious.
• High key mostly bright, minimal shadow. Clean, optimistic.
• Volumetric lighting visible rays through fog or dust. Spiritual, epic.
• Rim light edge light that separates subject from background. Essential for dark scenes.
• Softbox large diffused studio source. Clean, even, commercial.
• Neon saturated colored sign light, usually pink and cyan. Urban or retro futuristic.

The camera axis

Camera keywords shape how space is arranged and where the viewer stands. They are precise because they map to real photographic decisions the model has seen millions of times.

• Extreme wide shot subject tiny in vast environment. Scale and isolation.
• Wide shot full subject with context. The establishing frame.
• Medium shot waist up. The natural conversational default for people.
• Close shot head and shoulders. Emotion and expression.
• Extreme close shot a detail. Texture and intimacy.
• Eye level neutral and respectful. The honest default.
• Low angle looking up. Heroic, dominant, threatening.
• High angle looking down. Vulnerable, observed.
• Bird's eye view directly overhead. Pattern and geometry.
• Wide angle lens expands space, dramatic perspective. 16mm to 35mm.
• Normal lens natural perspective. Around 50mm.
• Telephoto lens compresses space, flatters faces. 85mm to 200mm.
• Shallow depth of field subject sharp, background blurred. Isolates the subject.
• Deep focus everything sharp front to back. For environments and landscapes.

The style axis

Style keywords decide whether your image is a photograph, a painting, a render, or an illustration, and within that, which tradition it belongs to. They are covered in depth in the styles article. Here is the operational set.

• Documentary photography honest, unposed realism.
• Editorial photograph styled, magazine quality.
• Cinematic film still graded, anamorphic, narrative.
• Oil painting thick impasto, rich color, visible brushwork.
• Watercolor transparent washes, soft blooms, paper texture.
• Gouache matte, flat, velvety. Beloved by illustrators.
• Ink wash painting stark, calligraphic, minimal.
• Flat vector illustration clean, geometric, modern.
• Concept art production oriented, environmental, detailed.
• 90s anime cel cel shaded, line driven, nostalgic.
• Isometric 3D render clean, geometric, playful.
• Clay render untextured gray material for form study.

The mood axis

Mood words are the loosest axis and the easiest to overuse. They work best when they name a specific atmosphere rather than a generic feeling. Somber is specific enough to move the model. Nice is not. Use mood words sparingly to reinforce what your light and style already imply.

• Somber quiet, heavy, serious. Pairs with low key and cool light.
• Ethereal light, delicate, otherworldly. Pairs with soft light and pale color.
• Gritty raw, textured, harsh. Pairs with hard light and documentary style.
• Serene calm, peaceful, balanced. Pairs with soft light and open space.
• Tense uneasy, charged. Pairs with mixed light and tight framing.
• Nostalgic warm, faded, memory like. Pairs with film stock and grain.
• Epic grand, heroic, vast. Pairs with wide shots and volumetric light.

The quality axis, and why you should mostly ignore it

This is the axis everyone overuses and that does the least. Words like 8k, highly detailed, masterpiece, trending on artstation, award winning, ultra realistic, intricate detail were genuinely useful on early diffusion models, which needed encouragement to produce clean output. Modern models do not. They already default to high quality, and these words have become filler that competes for attention without shaping the image.

The test is simple. Generate an image with and without a stack of quality words. On a current model you will usually see no meaningful difference, or even a slightly worse result because the attention budget was spent on filler. Cut them. Spend those words on light, camera, or style instead, where they actually change the output.

Combinations that reliably work

Some axis combinations have a track record of producing strong, coherent images because the choices reinforce each other rather than fight. Learning a handful of these proven combinations gives you a fast path to a finished look without rebuilding from scratch every time. Here are five worth memorizing.

The documentary portrait combines documentary photography, soft natural side light, a medium close shot on a 50mm lens, and a somber or quiet mood. It produces honest, unposed feeling portraits that read as real. The light and the style agree, which is why it holds together.

The cinematic hero combines cinematic film still, dramatic low key or volumetric light, a low angle wide shot, and an epic mood. It produces the blockbuster poster look. The low angle and the volumetric light both push toward grandeur, so they compound instead of competing.

The clean product combines studio product photography or 3D render, soft even softbox light, an eye level or slightly high angle, and a serene or neutral mood. It produces the ecommerce and advertising clean look. Every axis serves clarity, which is exactly what product work needs.

The editorial illustration combines flat vector illustration or gouache, a limited palette, a medium shot, and a nostalgic or wry mood. It produces the modern magazine illustration look. The medium and the mood align toward a graphic, considered sensibility.

The atmospheric landscape combines documentary or fine art photography, golden hour or volumetric light, a wide or extreme wide shot, and a serene or epic mood. It produces the sweeping nature image. The wide shot and the directional light both reward a vast environment.

Combinations that fight each other

Just as important is knowing which combinations sabotage themselves. When two words on the same axis push in opposite directions, the model averages them and the result loses definition. Avoid these unless you genuinely want a muddy middle ground.

Golden hour and blue hour describe light at opposite ends of the day with opposite color temperatures. Stacking them gives you a confused, desaturated wash. Pick the one that matches your mood.

Low key and high key describe opposite tonal ranges. One is mostly dark, the other mostly bright. Together they cancel into a flat mid tone image with no mood at all.

Shallow depth of field and deep focus describe opposite focus behaviors. Together the model picks one and ignores the other, or produces an inconsistent focus that reads as an error.

Wide angle lens and telephoto lens describe opposite perspective behaviors. Stacking them gives you spatially confused images where the perspective does not match the compression.

The rule underneath all of these is the same: one strong choice per axis. Two contradictory strong choices produce a weaker result than one clear one. When you are tempted to hedge by including both, you are usually better off committing to one and adjusting later if needed.

How to assemble a prompt from the axes

The framework becomes a method when you assemble prompts one axis at a time. Walk the axes in order, pick one or two words from each, and join them around your subject. Here is the assembly in slow motion, for a brief that says a chef in a kitchen.

1. Subject a chef in his fifties, tired eyes, plating a dish.
2. Setting in a small stainless steel kitchen, late at night.
3. Light warm practical light from a single overhead lamp, cool spill from a refrigerator glow.
4. Camera medium close shot, 50mm lens, shallow depth of field.
5. Style documentary photography.
6. Mood quiet dedication.

Strung together, that is a complete image brief built from the axes. Each word earned its place by controlling a specific dial. There is not a single quality superlative in it. And it will produce an image that looks like a decision was made, because one was, on every axis.

Building your own list

The list above is a starting point, but the most valuable cheat sheet is the one you build yourself from results you have actually seen. Every time a keyword moves an image in a direction you like, write it down under the correct axis. Every time a word does nothing, cross it off. Within a few weeks you will have a personal cheat sheet calibrated to your taste and your model, where every word is proven to work for you.

The point of a cheat sheet is not to have more words. It is to have the right words, understood well enough that you reach for them on purpose. Treat the axes as a menu of dials, pick deliberately from each, and your prompts stop being hopeful and start being controlled. That is the entire game, and no list of magic words will ever substitute for understanding which dial you are turning.