Did AI Staging Change Your Paint? The SWATCH Color Check

To check whether an AI staging tool changed your paint, flooring or cabinet colors, put the original and the staged file side by side at the same zoom, sample the same spots on each with a free eyedropper, and compare the RGB or hex values. Our rule of thumb: a gap of 6 points or less per channel (out of 255) is file noise, and a gap over 15 points on any channel, or channels moving in opposite directions, is a color shift worth a re-do. The SWATCH check below runs in about five minutes per photo with tools you already have.

Why does a color shift matter on a staged listing photo?

Furniture is the part of a virtually staged photo everyone knows is fake. Paint, flooring, counters and cabinets are not. They belong to the house, they convey with the sale, and a buyer uses the photos to judge them. If greige walls come back cool gray, or honey oak floors come back a fashionable ash, the photo now shows a house that doesn't exist.

The NAR Code of Ethics asks Realtors, in Article 12, to present a true picture in their advertising and other representations. A repaint nobody asked for is not a true picture, even when it looks better.

California adds a practical reason. Business and Professions Code Section 10140.8, the AB 723 rule in force since January 1, 2026, requires a disclosure that a listing image was altered plus a link, URL or QR code to the unaltered original. Any buyer can open both versions and compare the kitchen. A color mismatch that would have surfaced at the showing now surfaces on their phone. Many MLSs have their own altered-photo rules too. This isn't legal advice; check your MLS and your broker.

One more thing to settle first: compare against the exact file you will publish as the unaltered original. If your photographer delivered an edited HDR file, that file is your baseline, not a phone snapshot from the walkthrough.

What do you need to run a color check?

Nothing you have to buy. Any one of these reads pixel values:

• GIMP (free, Mac, Windows, Linux). The Color Picker tool shows RGB and hex in an info window. Turn on "Sample average" and set a radius of 2 to 5 pixels so you read a small patch instead of one noisy pixel.
• Photopea (free, runs in a browser). The eyedropper sets the foreground color; click the swatch to read the hex code. Set the sample size to an average if the option is offered.
• Digital Color Meter on a Mac (built in, under Applications, Utilities) reads the RGB value under your cursor on anything on screen.
• PowerToys Color Picker on Windows (free from Microsoft) does the same with a keyboard shortcut.

You also need somewhere to write six to ten pairs of numbers.

A word on the numbers. Each pixel has three values, red, green and blue, from 0 to 255. Hex is the same three values written in base 16, so #C8BEB0 is red 200, green 190, blue 176. If two samples differ, look at which channel moved and in which direction. That tells you more than the size of the gap.

The SWATCH check, step by step

SWATCH is our name for it. Six steps, one photo at a time, original and staged file.

1. S: Same view. Open both files in two windows at the same zoom, 100% if your screen allows, or both at "fit to screen". Line them up so the same wall sits in the same place on screen. Glance back and forth for ten seconds before you sample anything. Your eye catches a warm-to-cool swing faster than a number does, and it tells you where to aim the eyedropper.

2. W: White reference. Sample the ceiling and a run of painted trim, such as a door casing or baseboard, in both files. These are your control. Write down the values. If the trim moved by about the same amount and in the same direction as everything else you sample later, the whole image got a tone or white balance change. If the trim held steady and the walls moved, something repainted the walls.

3. A: Average the walls. Pick two spots on each painted wall you can see: one in direct light, one in shade. Stay at least a hand's width away from any new furniture, rug or lamp, because added pieces cast shadows and bounce light that legitimately change nearby pixels. Use the same spots in both files. Sample with the averaging radius on.

4. T: Tone of the floor. Sample bare floor in two places, one near the camera and one deeper in the room. If the staging added a rug, sample the exposed floor beside it, not the strip right at its edge where a new shadow falls. On wood, watch the red channel: orange and honey tones carry a high red value, and a drop in red with blue holding steady is the signature of a floor going gray. On tile or carpet, watch for any channel moving more than the others.

5. C: Cabinets and counters. Sample the center of one cabinet door face, away from the hardware and the edge bevels, then a flat stretch of countertop and backsplash. Look at the hardware too. Brushed brass turning to nickel is a finish change even if the cabinet paint held. In bathrooms, do the vanity and the tile surround.

6. H: Hex log and decide. Put each pair side by side and apply the table below. One flagged sample on a surface that matters is enough to send the photo back for a re-do. If the shift is global (Step 2 says the trim moved too), a simple re-export or a re-do usually fixes it. If it is local, the tool changed that surface, and you should not publish that version.

Where to sample and what counts as a shift

The thresholds here are this site's rule of thumb, not an industry standard. No real-estate body publishes a color tolerance for staged photos. We set the floor at 6 points because JPEG compression and resizing alone routinely move values by a few points on flat surfaces, and the ceiling at 15 points because a gap that size on a flat painted surface sits comfortably above the just noticeable difference described below. Tighten it if your own checks suggest you should.

Between 7 and 15 points is a gray zone. Look again with your eyes, check whether the trim moved the same way, and decide whether a buyer standing in that room would say the color is different. If you would hesitate to defend it, re-do it.

If you want a more formal number, color science measures differences as Delta E, a distance between two colors in a space designed to match how people see. The Wikipedia article on color difference explains the formulas and notes that a CIE76 Delta E of about 2.3 corresponds to a just noticeable difference. Free online Delta E calculators take two hex codes. That is a finer tool than most listings need, but it is there if a seller disputes a result.

What usually causes a color shift?

Three things. Only the last one means a surface of the house was redrawn.

• A global tone change. The whole image comes back warmer, cooler, brighter or more saturated. Trim and ceiling move with everything else. This can come from the staging step or from an export setting along the way, such as a color profile being dropped or converted. Confirm your own editor exported the original with an embedded profile before you blame the tool.
• Lighting the staging added. New lamps, a rug or a sofa against a wall add shadow and bounce. A wall can read darker right behind a new sofa, and that is fine. That is why the SWATCH steps sample away from new pieces.
• A surface the tool redrew. The walls, floor or cabinets come back a different color while the trim and ceiling hold. This is the one that misrepresents the property. It is also the hardest to see in a quick look, because the new color is usually plausible and often more fashionable than the original.

How do the tools compare on color?

We haven't run a documented test across staging tools, so we won't rank anyone here. What we can describe is what each tool says it is built to do, and that is worth checking before you upload.

QuickStage is built to keep walls, floors, windows and paint colors exactly as photographed, and to add only movable furniture and decor. That design goal is the right one for color, but "built to" is not a promise about every photo, so run SWATCH on its output the same way you would on any tool's. If a photo fails, QuickFix re-dos are included on plans. Pay-as-you-go is $1 per photo with no subscription, plans run as low as $0.11 a photo on the annual Agency tier, and results come back in about 30 seconds, so a re-do doesn't turn into a budget question or a lost afternoon.

For any other tool, look for a plain statement in its own documentation about whether it alters existing surfaces. If you can't find one, assume nothing and let the eyedropper decide.

Bottom line

Paint, floors and cabinets are part of the house, so a staged photo has to show them as they are. Open both files at the same zoom, sample the same spots with a free eyedropper, use trim and ceiling as your control, and flag anything over 15 points on a channel or any warm-to-cool swing on a surface. Five minutes per photo is cheap next to a buyer comparing your staged kitchen with the original you are required to link in California.

Run the SWATCH check on three free QuickStage photos, no account or card needed.