Buyers decide in about two seconds, and they decide on the thumbnail. Drop 3 photos from a live listing below — MLS, Airbnb or rental — and each one comes back graded, with the specific thing wrong with it, what that costs, and how to fix it. No account, no card, and the files are read once and discarded.
Nothing is stored. The photos are read once, scored in your browser, and never saved to an account you don't have.
A vision model reads each photo and reports signals — how bright it is, how cluttered, whether the vertical lines are vertical, whether the windows are blown out, whether anything identifying the occupant is visible. It never issues a verdict; the arithmetic that turns those signals into a grade runs in your browser against a fixed, published table of faults, which is why the score moves the instant you switch platform and why two identical photos can never grade differently.
Coverage is the finding agents underrate most. A missing room is not neutral — buyers assume the room you didn't photograph is the bad one, and they assume it for free. These are the 7 shots the audit checks a set against, weighted by how much their absence costs.
Coverage and running order are questions about a whole set, so they need a whole set — which is why they sit behind a free account rather than being answered badly from three photos.
An MLS listing is browsed by someone who has already filtered by price and postcode; an Airbnb listing is browsed by someone choosing between forty properties in one scroll. The cover photo does the click-through and the next four do the booking. Nobody reaches photo twenty, which means the ordering matters more on Airbnb than anywhere else and the tail matters less.
The practical differences: lead with the space that makes the stay, not the front of the building. Shoot in the light guests will actually see — a sunset shot of a west-facing balcony is fair, a night exterior is not. Show the bed, the bathroom, the kitchen and where people sit, because a guest who cannot find one of those assumes it is grim. And keep every photo true: an MLS photo oversells a house someone will visit, an Airbnb photo oversells a room someone will wake up in, and the second one comes back as a refund.
Set the platform above to Airbnb and the guidance changes with it. The faults are the same everywhere — dark is dark — but what a complete set looks like is not.
Almost every fault this tool finds is a shooting fault, not an editing one, which means almost every one of them is free to avoid. A modern phone has more than enough sensor for a listing photo. What it does not do is choose the hour, the height or the corner for you.
Shoot interiors mid-morning or late afternoon, when light comes through the windows at an angle rather than straight down. Avoid midday in summer: the sun is directly overhead, the windows blow out to white, and the room behind them goes muddy. For the exterior, work out which way the front faces and shoot when the sun is on it — a north-facing façade photographed at noon is in its own shadow all day and will never look like anything.
Every lamp, every overhead, every under-cabinet strip, in every room, even in daylight. It is counter-intuitive and it is the single biggest improvement available to most sets. Interior lamps fill the shadows the window cannot reach and give the room warmth the camera reads as lived-in. If a bulb is dead, replace it before you shoot; if two bulbs in one fitting are different colours, the photo will show it even though your eye did not.
A shot taken from the doorway flat at the opposite wall flattens the room and makes it look smaller than it is. Back into a corner instead so two walls and a run of floor are visible. Depth is what tells a viewer how big a room is; a single wall tells them nothing. Hold the phone at chest height, roughly four to five feet, and keep it vertically level — tilting up stretches the ceiling, tilting down makes the floor swallow the frame, and both produce the leaning walls the analyzer flags.
Every platform crops to a landscape thumbnail, so shoot landscape and stop losing the top and bottom of your own frame. Resist the ultra-wide lens: it exaggerates the room in a way viewers have learned to distrust, and it bends straight lines at the edges. The standard lens, from a corner, shows a room honestly — and honesty is what stops a viewing turning into a complaint on the doorstep.
Faults are universal, but what a viewer is looking for changes with the room. These are the things worth checking before you press the shutter.
Clear the worktops completely — kettle, toaster, knife block, the lot — then leave one deliberate object, a bowl of fruit or a board, so it reads as a kitchen rather than a showroom. Remove everything from the fridge door. Shoot from the corner that shows the run of units and the window together, and include the floor: buyers read the flooring as a proxy for how recently the kitchen was done.
This is usually the cover photo, so it gets the most care. Straighten the cushions, square the rug to the room, and get the cables out of frame. Turn the television off — a black rectangle is a hole in the middle of the picture, and a paused film dates the listing. If there is a fireplace or a bay, compose so it is the anchor rather than something clipped at the edge.
Make the bed properly, including the far side nobody can see from the door, because the camera can. Clear the bedside tables to a lamp and nothing else. Photograph from the corner diagonally opposite the bed so the room has depth, and include the wardrobe or the window — storage and light are the two things a bedroom is judged on.
Everything personal goes: toothbrushes, bottles, towels in use. Put out clean towels, close the toilet lid, and dry the glass and taps — water spots and smears are the detail that makes a clean bathroom photograph dirty. Watch your own reflection in the mirror, which is the most common reshoot in the whole set.
Move the bins and the cars, including yours. Shoot from an angle rather than dead-on so the building has some form, and crop out the neighbour's skip if you can do it without cutting the roofline. Nothing dates a listing faster than the wrong season in the garden, so if the photo was taken in November and it is now May, reshoot it — the analyzer cannot see the calendar, but every buyer can.
Set size and running order are separate decisions and people routinely get one right and the other wrong. A large, badly-ordered set performs worse than a small, well-ordered one, because attention runs out long before the photos do.
For an MLS listing, aim for twenty to thirty and lead with the exterior — the buyer is confirming a house they have already shortlisted on price and location, and the front elevation is how they recognise it. Follow with the main living space, the kitchen, then bedrooms in descending size, bathrooms, and finally garden, garage and any outbuildings. Floor plans belong near the end where they answer questions rather than raise them.
For a short-term rental, aim for fifteen to twenty-five and invert the opening. The guest is not confirming a known property; they are choosing between dozens in one scroll, so the cover photo has to be the most appealing space in the listing — the living area, the view, the pool, whatever makes the stay. The exterior can wait until the middle of the set.
For a long-term rental, ten to fifteen is enough, but every room needs one. Rental enquiries are dominated by people filtering for what a place lacks, and an unphotographed room reads as a lacking one.
Whatever the platform, the last photo matters more than the ones in the middle: it is what a viewer is looking at when they decide whether to enquire. Finish on something good rather than on the boiler cupboard.
Ask a general-purpose model whether a photo is good and you get a different answer on Tuesday than you got on Monday, with no way to tell which one to act on. This tool splits the job in two. The model is only allowed to report what it can see — brightness, clutter, whether verticals are vertical, whether windows are blown out, whether anything identifying the occupant is in frame. Turning those observations into a number is done afterwards, in your browser, against the published fault table above.
Two consequences follow, and both are the point. The same photo always grades the same, so a reshoot that scores higher really is better rather than differently judged. And when a threshold is sharpened, every audit ever saved re-derives under the new rule instead of being frozen at whatever the model happened to say the day it ran.
It also means you can argue with it. Every deduction is named and priced, so if you disagree that a bright room is overexposed you can see exactly which line you are disagreeing with — rather than being handed a score of 61 and left to guess.
For an MLS listing, 20 to 30. For an Airbnb, 15 to 25 with the best five first, because nobody scrolls to twenty on a phone. For a long-term rental, 10 to 15. Below about ten photos on any platform, buyers and guests start assuming something is being hidden — which is why the score docks a thin set even when every individual photo is good.
The faults that cost the most are the boring ones: the room is too dark, the lights were left off, the windows are blown out to white, the camera was tilted so the walls lean, personal items are visible, or the shot was taken flat at one wall so the room has no depth. This tool checks 11 of them per photo and tells you which one to fix first, ranked by the points it gets back.
It is most of the decision. On a search results page the first photo is the entire listing — the price and the address are text next to a thumbnail, and the thumbnail is what gets clicked or scrolled past. On Airbnb the cover photo drives click-through and the next four drive the booking. If your best photo is sitting at position seven, reordering costs nothing and is usually the single highest-return change available.
Shoot in daylight with every lamp switched on, from a corner rather than a doorway so two walls and the floor are visible, holding the phone level so the verticals stay vertical. Then put the most appealing space first — for a short-term rental that is usually the living space or the view, not the exterior an agent would lead with. Set the platform to Airbnb above and the analyzer grades the set against those rules instead of the MLS ones.
Yes. function(){throw Error("Attempted to call DEMO_PHOTO_LIMIT() from the server but DEMO_PHOTO_LIMIT is on the client. It's not possible to invoke a client function from the server, it can only be rendered as a Component or passed to props of a Client Component.")} photos are graded here with no account, no card and nothing stored — the images are read once and discarded. A free account raises it to five photos per audit and unlocks the listing-level findings: coverage against the shots buyers expect, the running order, and the reshoot list ranked by points recovered.
No. On this page the file is read, graded and deleted in the same request — there is no account to attach it to and nothing is written to a database. The scoring itself happens in your browser, which is also why the grade changes instantly when you change platform.
A vacant room grades fine and still sells badly, because a buyer has nothing to judge the scale against. Virtual staging furnishes it without touching the floors, walls or fixtures — and shows you the whole plan before you make an account.
Stage an empty room free