7 Tests to Run on Any AI Virtual Staging Tool Before You Pay

Before you pay for an AI virtual staging tool, run three of your own empty-room photos through its free trial and score seven tests: structure, invented fixtures, clear walkways, furniture scale, resolution and watermarks, true cost per usable photo, and fine print. Any failure on the first three is a veto, not a deduction. Survivors get compared on cost per usable photo (price per photo ÷ the share of outputs you'd actually publish), never on sticker price.

We call that "veto, then divide." Here's how to run it, with a scorecard to copy.

How do you set up a fair free trial?

Three photos are enough if they stress the tool. Use the same three on every tool, and keep the untouched originals for side-by-side checks (and, in California, for the disclosure).

1. The window room. Two or more windows, a wall color that isn't plain white, lots of visible floor. It carries Tests 1 and 4.
2. The doorway room. An open doorway, hallway or sliding door in frame, so there's a walking path to block. It carries Test 3.
3. The bare-wall room. A long empty wall or alcove where a tool might "helpfully" add a fireplace, shelving or an appliance. It carries Test 2.

Tests 1–4: Does the staged photo still show the real room?

Test 1: Are the windows, walls and floors unchanged?

Put the original and the staged image side by side at full size:

• Windows. Same count, position and pane grids.
• Doors and fixtures. Doors, vents and ceiling lights stay put. Hidden behind a new sofa is fine; vanished from open wall is not.
• Wall color. Eyedropper the same patch of wall in both images with any photo editor. A slight brightness shift can be relighting; a different color family (beige to gray, white to greige) is a repaint.
• Floors and trim. Same flooring, plank direction and tile pattern; same beams, baseboards and molding.

Score 2 if everything matches, 1 for a small artifact a re-do could fix, 0 if a window, wall, floor or paint color changed. It's a veto because buyers tour the real room. In California, Business and Professions Code §10140.8 (AB 723), in force since January 1, 2026, requires digitally altered listing images to carry a disclosure plus a link, URL or QR code to the unaltered original, so any structural change sits one click from the evidence. This isn't legal advice; check your MLS and broker.

Test 2: Did it invent built-ins or appliances?

Apply the contractor rule: if adding it in real life would take a contractor or an appliance delivery, it must already be in the original. Check every output for:

• fireplaces, mantels and built-in shelving;
• kitchen islands, range hoods and any appliance;
• ceiling fans, pendants and recessed lights;
• wall paneling, crown molding and installed blinds.

Sofas, rugs, lamps and art are fair game; a wall-mounted TV is a judgment call. Score 2 if nothing on that list appeared and 0 if anything did; there's no partial credit. Article 12 of the NAR Code of Ethics requires REALTORS® to present "a true picture" in their advertising, and a dishwasher the house doesn't have is hard to square with that.

Test 3: Are doorways and walkways still clear?

On the doorway room, trace a finger from each doorway to the middle of the room and on to the next opening. If your finger crosses furniture, the tool failed. Also check that sliding and French doors, closets and stairs aren't blocked. A sofa across a doorway tells buyers the room doesn't work.

Score 2 for clear paths, 1 for a tight squeeze, 0 for any blocked door.

Test 4: Is the furniture the right size and sitting on the floor?

Use a door as a ruler, comparing pieces about as far from the camera as the door. A standard U.S. interior door is 80 inches tall, so a sofa seat (about 17–19 inches) should land a little under a quarter of the way up it, and a dining table top (about 28–30 inches) a bit over a third. Then check:

• Contact. Legs meet the floor with a soft shadow. Nothing floats or sinks into carpet.
• Vanishing lines. Extend a floorboard or tile line; parallel rug and bed edges should head toward the same point.
• Light. Shadows fall away from the windows, matching the room's existing shadows.

Score 2 if every piece looks photographed in place, 1 if one is off, 0 if the room reads as a collage.

Tests 5–7: What will a usable photo really cost?

Test 5: What does the free output actually give you?

Download the file (not the browser preview) and check:

1. Pixel size. On a Mac, Preview → Tools → Show Inspector; on Windows, right-click → Properties → Details. Compare with your original.
2. Watermarks. Corners and center, including faint, semi-transparent text.
3. Compression. At 200% zoom where furniture meets a bright window, look for halos, smeared wood grain and blocky shadows.

A watermark on a full-size file is fine; you can still judge quality. A downsized file isn't a test: you can't see what you'd pay for. Score 2 for full-size and clean, 1 for full-size with a watermark, 0 for a shrunken preview. Then confirm the paid file meets your MLS's photo specs.

Test 6: What's the true cost per usable photo?

Sticker price assumes every output is a keeper. Run this instead:

True cost per usable photo = price per staged photo ÷ usable rate

Usable rate is the share of outputs passing Tests 1–4 without a re-do. If re-dos bill like new photos, you pay for 1 ÷ usable rate attempts per keeper on average.

For example, stage the same 10-photo listing on two tools, paying for every re-do:

• Tool A: $0.40 per photo, 50% usable. $0.40 ÷ 0.50 = $0.80 per usable photo.
• Tool B: $0.60 per photo, 90% usable. $0.60 ÷ 0.90 = about $0.67.

The long way agrees: 10 keepers at a 50% hit rate take about 20 attempts, and 20 × $0.40 = $8.00, or $0.80 each. The lower sticker costs 20% more per photo you can publish. Two cautions: three trial photos are a small sample, so recheck your rate on the first paid batch; and when re-dos are free, a low usable rate costs time instead of money.

Test 7: What's hiding in the fine print?

Before entering a card, answer four questions:

1. Do credits expire? Find the window, the rollover rules, and what happens on cancel or downgrade. An expired credit is a photo with a usable rate of zero.
2. Per credit or per photo? How many credits does one finished photo cost, and do high-resolution downloads, extra styles or re-dos cost more? Convert first: at two $0.40 credits per photo, Tool A's sticker becomes $0.80 and its true cost $1.60.
3. Is the headline rate an annual prepay? Some tools quote their lowest rate only for a year paid up front. Divide that payment by the photos you'll really stage in 12 months, not the plan's allowance.
4. What does the license allow? Check commercial use (MLS, portals, print, ads), whether rights survive cancellation, and what the tool may do with your uploads. If a photographer shot the listing, confirm their license allows edits.

Score 2 if the answers are easy to find, 1 if they're buried in the terms, 0 if they're missing or would strand money you've paid.

How do you score the results?

Copy this, one column per tool:

2 means pass, 1 minor or fixable, 0 fail; the six scored tests max out at 12. Veto: any 0 on Tests 1–3 disqualifies a tool, whatever its total. Divide: among the survivors, the lowest true cost per usable photo wins, and the other scores break ties.

Which tool lets you run all seven tests free?

Any tool with a real free trial can be scored this way, and some will score well. What matters is whether you can finish the scorecard before handing over a card.

With QuickStage you can: 3 free photos with no account or card, results in about 30 seconds, exactly the setup above. Its design targets Tests 1–3. It's built to keep walls, windows, floors and paint colors exactly as photographed, add only movable furniture and decor, never add appliances or built-ins that aren't there, and keep doors and walkways clear. It aims for top-quality, photorealistic results; Test 4 checks that. Don't take that on faith; score it like any other tool.

Its numbers for Tests 6 and 7: $1 per photo pay-as-you-go with no subscription; monthly plans from $19.99, as low as $0.13 per photo; and $0.11 per photo on the annual Agency plan, billed yearly. One credit is one staged photo; pay-as-you-go credits don't expire, and plan credits reset monthly. QuickFix re-dos are included on plans, up to a monthly allowance, without using a staging credit. The $0.11 rate is an annual prepay, so give it the Test 7 check; lighter users should compare $1 pay-as-you-go or the $19.99 plan.

Bottom line

Three photos, seven tests, two passes: veto anything that changes the room, invents fixtures or blocks a doorway, then divide price per photo by usable rate for whatever survives. A tool that fails Tests 1–3 isn't cheap at any price. Read the fine print before the card comes out, and whichever tool wins, label staged photos and keep the originals.

Try three photos free on QuickStage, no account or card needed, and fill in the scorecard before you pay anyone.