Quick answer

Product photography lighting is not a mood word you sprinkle into a prompt. It is a set of four decisions you can describe in plain language: where the light comes from (direction), how big the light source is relative to the product (softness), what the light does when it hits a shiny surface (reflections), and how the product meets the ground (contact shadows). Add a fifth decision — consistency — and you have a brief an image model can actually follow.

If you only remember one rule: describe the light the way you would describe it to a retoucher who cannot see your reference photo. "Soft light from the left" is weak. "Large diffused source at 45° camera-left, slightly above product height, so the shadow falls to the right and the front label stays evenly lit" is a brief.

Why lighting language beats style adjectives

Most weak product prompts stack adjectives: premium, clean, minimal, studio, high-end. Those words describe a feeling, not a setup. Two people reading "premium studio lighting" will picture completely different images, and so will two models.

Lighting language works better because it is spatial and physical. Direction, source size, and surface behavior are things you can point at. When a result comes back wrong, you can name which variable failed and change only that one — instead of rewriting the whole prompt and hoping.

This is also why a lighting reference is portable. The same four-variable brief works whether you are generating a hero shot, a lifestyle scene, or a packaging mockup. Only the variables change.

The five variables, translated

1. Direction: where the light sits

Direction is the single highest-leverage variable. State it as a clock position or an angle relative to the camera, and say whether the source is above or below product height.

Direction	What it does to the product	Good for	Watch out for
Front (on-axis)	Flattens texture, even coverage	Catalog-style clarity, label legibility	Looks lifeless; no shape definition
45° camera-left or right	Classic shape + shadow balance	Most hero shots	Shadow can crowd the frame edge
Side (90°)	Strong texture and edge definition	Matte, textured, or fabric surfaces	Half the product goes dark
Back / rim	Separates product from background	Glass, bottles, dark-on-dark	Front needs a fill or it goes muddy
Top-down	Even, overhead, "flat lay" feel	Flat objects, packaging tops	Weak vertical form; hard shadows under edges

Write direction as a sentence fragment, not a vibe: key light 45° camera-left, slightly above product height.

2. Softness: source size, not "soft" as a mood

Softness comes from the size of the light source relative to the product. A big source close to the product gives gradual shadow edges and smooth highlights. A small or distant source gives crisp shadow edges and a hot specular highlight.

Useful phrasings:

large diffused source, close to product → gradual falloff, gentle highlights
small hard source, distant → defined shadow edges, punchy specular
soft top light with subtle fill from camera-right → even, low-contrast, editorial

Avoid the bare word "soft" alone. Pair it with a size cue: soft, from a large source.

3. Reflections: what shiny surfaces do

Reflections are where AI product images most often fall apart. A glossy bottle, a chrome cap, or a screen will mirror whatever the model imagines around it. You have three options, and you should pick one deliberately:

Suppress them: matte finish, no visible reflections, no mirror-like highlights
Control them: long vertical highlight strip down the left side of the bottle, single clean reflection, no clutter reflected
Use them: mirrored reflection on a polished dark surface beneath the product

For glass and metal, also state the background, because the product reflects it: against a plain mid-grey background so reflections stay clean.

4. Contact shadows: how the product touches the ground

Contact shadow is the small dark area where the product meets the surface. It is the fastest way to make a generated product look like it is floating. Say it explicitly:

tight contact shadow directly under the base, soft edges
product sits flat on the surface, no floating, no gap under the base
long soft shadow cast to camera-right, consistent with key light at 45° camera-left

The last one matters most: the cast shadow direction must match the key light direction. If the light is camera-left, the shadow falls camera-right. Mismatched shadows are the most common tell in AI product imagery.

5. Consistency: the same brief across a set

If you are generating a set — five SKUs, three angles, a carousel — consistency is its own variable. Lock the lighting block as a fixed string and change only the product and angle. Keep the same direction, source size, background, and shadow language in every prompt. Changing one word like "soft" to "dramatic" between frames will visibly break the set.

Step-by-step workflow

Pick one reference behavior, not a reference photo. Decide the single lighting idea you want: "big soft source, clean vertical highlight, tight contact shadow."
Write the lighting block first, before the product description. Four lines: direction, softness, reflections, contact shadow.
Add the product and background after the lighting block, so lighting stays the dominant instruction.
Generate two or three variants that change exactly one variable — usually direction or source size.
Review against the checklist below and note which variable failed.
Freeze the winning lighting block as a reusable string for the rest of the set.
Iterate on the product, not the light, once the lighting block is locked.

You can run this loop in the AI image generator workspace, where you can keep a prompt and its variants side by side and compare results before committing to a set. Availability of specific models and the credit cost of each generation are shown in the live model selector and the quote displayed at generation time — check there rather than assuming a given model is available.

Example input (copyable)

This is an illustrative prompt, not a tested result. It shows the structure: lighting block first, product second.

text

Studio product photograph of a 250ml amber glass dropper bottle, unbranded.
Lighting: key light 45° camera-left, slightly above product height, from a large
diffused source close to the product. Soft gradual shadow edges. Single clean
vertical highlight strip down the left side of the glass; no clutter reflected.
Tight contact shadow directly under the base, cast shadow falling to camera-right.
Background: seamless mid-grey, no gradient banding.
Camera: eye-level, product centered, shallow depth of field, sharp on the label area.
No text, no logos, no hands.

Variant A — softer, more even

Change only the softness line: very large diffused source, front-left, with subtle fill from camera-right; minimal shadow contrast.

Variant B — harder, more directional

Change only the softness line: small hard source at 45° camera-left, distant; defined shadow edges, crisp specular highlight.

Variant C — backlit glass

Change only the direction line: strong rim light behind the bottle, soft fill from camera-front; bright edge separation, dark clean background.

Three prompts, one variable each. That is what makes the comparison readable.

Output review checklist

Judge these subjectively — they are visual judgments, not measurements:

Shadow direction matches key light. Light camera-left, shadow camera-right.
Contact shadow exists and is tight under the base. No floating product.
Highlight shape is intentional. A clean strip, not a random blob or a reflected room.
Reflections contain nothing you did not ask for — no phantom windows, hands, or text.
Label and edges stay legible where the light is brightest.
Background is clean with no banding or unintended gradient.
Set consistency: across variants, direction and shadow language are unchanged.
No invented branding or text on the product.

If a check fails, change the matching line in the lighting block only. Do not rewrite the whole prompt.

Common mistakes

Adjective stacking. "Premium, cinematic, luxurious lighting" gives the model nothing spatial to work with.
Omitting contact shadow. The product floats, and no amount of "realistic" fixes it.
Contradicting yourself. Asking for a hard shadow and a soft diffused source in the same prompt.
Ignoring reflections on glossy products. Glass and chrome will reflect whatever the model invents unless you constrain it.
Changing the lighting between frames of a set. Consistency breaks the moment one prompt says "dramatic."
Treating parameters as universal. Provider-specific parameters are not portable across models — a flag that works for one video or image model may be ignored or rejected by another. Write lighting in plain language and keep it model-agnostic.
Confusing a lighting brief with a camera-buying decision. You do not need equipment knowledge to write this brief; you need spatial language.

FAQ

Do I need to know real lighting gear to write these prompts? No. You need to describe direction, source size, reflection behavior, and contact shadow. Gear names are optional shorthand, not requirements.

What if the model ignores my lighting block? Move it earlier in the prompt and shorten the rest. Long product descriptions dilute lighting instructions. Also check for contradictions between lines.

How do I keep a set consistent? Freeze the lighting block as an exact string and reuse it verbatim. Change only the product, angle, or background line.

Can I use the same lighting language for video? The vocabulary transfers, but parameters do not. Provider-specific flags are not portable between models, so keep the brief in plain language and verify what the target model accepts.

Where do I check which models are available and what a generation costs? In the live model selector and the quote shown at generation time. Pricing for the workspace itself is on the pricing page.

Is a lighting reference the same as a style reference? No. A style reference says how the image should feel. A lighting reference says where the light is, how big it is, what it reflects, and where the shadow lands. The second one is actionable.

Product Photography Lighting as an AI Image Brief: Direction, Softness, Reflections, and Contact Shadows