MLS photo tagging and watermark checks: what happens to your listing photos after upload

What happens to a photo when you upload it

Uploading a photo to the MLS used to mean it was stored and shown. On many MLSs today, several automatic steps run on each photo as it arrives.

• Tagging. Software recognises the room type and the features it can see, and labels the photo.
• Captions. A short description of the photo is written into the photo's description field.
• Auto-filled listing fields. Features found in the photos are offered as values for the listing form.
• Compliance checks. The photo is checked for watermarks, logos, text, signs, people and duplicates.
• The MLS watermark. Some MLSs stamp their own mark on every photo after upload.

Which of these your MLS runs depends on its software and its contract with the vendor. The rest of this guide explains each step and what it means for the photos you upload.

Photo tagging and room recognition

Real estate photo recognition means software that looks at a listing photo and says what is in it. Restb.ai, the vendor behind many MLS tagging tools, says its image tagging models detect more than 500 different details, including room types, interior and exterior home features, and architectural styles.

It reaches agents through MLS software vendors. Restb.ai lists Cotality (formerly CoreLogic), dynaConnections, Rapattoni, FBS, ICE, Stratus Data Systems and NAVICA as partners. Three examples of how it shows up:

• Paragon. Paragon's 6.3 and 6.4 release notes describe an optional integration with Restb.ai that automatically applies RESO labels to all uploaded photos, plus photo descriptions. They also say users can review and modify the labels and descriptions.
• Rapattoni. In February 2023, RealTrends reported that Rapattoni expanded its Restb.ai partnership to photo tagging, ADA accessibility compliance, auto-population of listing details and enhanced search.
• dynaConnections. When bridgeMLS went live in April 2023, Restb.ai said its tags cover 50+ RESO-standardized room classifications and more than 100 RESO-standardized property features, and that detected features are presented to members for review and filled in once confirmed.

RESO is the Real Estate Standards Organization, which sets the shared field names MLSs use. Tags that follow its data dictionary mean a kitchen photo is labelled the same way in every MLS that uses them, and the tags can travel with the listing to other sites. Baldwin REALTORS' help page says the tags syndicate to feeds so consumers can find properties with specific features.

Portals build their own systems too. In March 2025, Built In described Realtor.com's on-demand image tagging service for MLS partners, which uses OpenAI's CLIP and Google's ViT models. The engineers' reason was that agents name the same home features differently in different regions, so searches and tags don't always match.

Captions, accessibility and auto-filled fields

A caption is the text description attached to a photo. Restb.ai says as few as 5% of MLS listings contain information at the photo level, and it pitches its captions as a way for sites to meet ADA and WCAG accessibility expectations: a screen reader can read a caption, but not a picture.

Auto-population is the step agents notice most. If the photos show a fireplace, a kitchen island or stainless steel appliances, the MLS can offer those as values in the listing form. Restb.ai's MLS page says it scans every image uploaded to the MLS and pre-populates the appropriate fields.

The tags are a suggestion, not a ruling. Paragon and bridgeMLS both describe the agent reviewing the result before it is used, so read what was filled in. Software can see a feature that doesn't convey, or miss one that is only in the listing remarks.

Condition and quality scores from photos

Can software tell the condition of a house from photos? Several vendors sell exactly that, and their customer lists lead with appraisers, lenders, investors and valuation companies.

• Restb.ai Property Condition scores the condition, quality and potential of homes from their property photos, with sub-scores for the kitchen, bathrooms, interior and exterior, and a confidence score. It offers a C1–C6 model that follows Fannie Mae's appraisal condition scale, a Q1–Q6 model for quality, and its own R1–R6 model. On the C scale, a higher number means worse condition. Its customer list includes AVMs, iBuyers and investors, appraisals and inspections, MLSs, property search and insurance companies.
• HelloData QualityScore says it analyzes real estate photos to assess the condition and quality of single-family and multifamily properties, rates curb appeal, and detects room types and features such as stone countertops. It names appraisers, investors, lenders, brokers, MLS and ILS providers and proptech companies as users.
• Proptexx lists condition, quality and compliance reviews among the uses of its property intelligence for portals, MLSs and agents.

In May 2026, Restb.ai wrote that Stewart Valuation Intelligence uses its condition and quality scores in its appraisal review process. None of these vendors publish how a score affects a specific listing, so treat a score as one more reader of your photos. A photo that shows a room clearly and evenly lit gives any reader, human or software, less room to guess.

The MLS watermark, and why you should never remove it

Some MLSs stamp every uploaded photo with their own mark. CRMLS, a California MLS, says in its 2026 Photographs and Media FAQs that each uploaded photo is stamped with a watermark showing the letters "CRMLS" in the bottom-right corner. CRMLS says the mark indicates that the photo is original and that it is part of the CRMLS compilation copyright.

That is how CRMLS catches reused photos. If a photo is copied from another listing and uploaded again, the old mark carries over and a new one is added, which CRMLS calls a "double watermark". CRMLS says reusing MLS photographs is prohibited even with the original owner's permission, that it violates rule 11.5(d), and that you should get the originals or unmarked versions from whoever grants permission.

HiCentral MLS in Hawaii announced in 2015 that it adds a watermark to the lower-right corner of each photo after upload. It says the mark lets its software find HiCentral listing data posted on unauthorized websites, and that it appears on photos in IDX and third-party feeds too.

Removing or covering an MLS watermark to reuse a photo works against exactly what the mark protects: the MLS's compilation copyright and the photographer's rights. Don't do it, and don't use watermark-removal tools on MLS photos. If you need old photos of a property, ask the listing agent or the photographer for the originals and a written license. Our guide on who owns listing photos covers the licensing.

Some MLSs also ask for a label on altered photos, such as a "Digitally Altered" or "Virtually Staged" mark. That is a disclosure, not the MLS's own watermark; see our guide to AI photo editing and MLS rules.

Watermark, logo and branding detection

Most MLSs ban branding inside listing photos. CRMLS defines branding as any media content that identifies a person or entity, and lists signs, contact information, identifiable persons, logos, slogans, catchphrases, external links, team names and anything else tied to the broker or agent. Its FAQ says a photo with a "For Sale" sign counts as branding, and that branding brings removal of the photo, a citation and an immediate fine with no warning notice.

Detection software looks for exactly these things. As the vendors describe their own products:

• Restb.ai Watermark Detection checks every uploaded image and video for artificially added watermarks, logos and text overlays.
• Restb.ai Photo Compliance flags people and faces, contact information, "for sale" signposts, religious or political objects, renders and virtual staging, watermarks, logos and added text, and exact and near duplicates. Restb.ai names Stellar MLS, Heartland MLS, Buffalo Niagara Association of Realtors and SWMLS among the MLSs using it.
• Proptexx lists checks for watermarks, logos and text; people, signs and contact details; renders and virtual staging; and duplicate images and listings, run against the customer's own rules and review queues.

A flag usually goes to the MLS's compliance staff or straight back to you, depending on how your MLS has set it up. Our MLS photo rules by MLS table shows which of 22 MLSs ban branding, people and signs.

Why listing photos get rejected, and a pre-upload checklist

Most rejected or flagged photos fall into a few groups: branding, people, photos you don't have the rights to, and files the MLS can't take. Run through this list before you upload:

1. No logos, watermarks or text. Take your name, logo, phone number, website and any photographer's watermark off the MLS version. Keep a branded copy for your own channels.
2. No signs. Shoot the exterior without the yard sign, or before it goes up. At CRMLS a "For Sale" sign counts as branding.
3. No people. Nobody in the frame, including your reflection in mirrors, windows and appliances.
4. Only photos you have the rights to. No photos copied from an old listing, Google or another website. CRMLS treats publication through the MLS as commercial use and asks for written permission from the photographer or seller.
5. Disclose any alteration your MLS asks you to. Virtual staging and AI edits usually need a label and sometimes the original next to it.
6. Check size and format. Each MLS sets its own pixel, file-size and photo-count limits. Our photo size guide lists them, and the free listing photo resizer resizes a batch in your browser.
7. Review the tags. After upload, read the room labels, captions and any auto-filled fields, and correct what is wrong.

How to shoot so the tags come out right

The tagger can only label what the photo shows. These habits help it, and they help buyers too:

• One clear photo per room that shows what the room is. The kitchen with its counters, cabinets and appliances. The bathroom with its sink, toilet and tub or shower. The bedroom with the bed. A close-up of a faucet tells software, and a buyer, very little about the room.
• Show the features you want found. If the fireplace, the island or the walk-in shower matters, make sure it is clearly in frame.
• Keep the camera level. Hold the phone or camera at about chest height with vertical lines straight. Tilted walls make a room harder to read.
• Light the room evenly. Turn on the lights, open the blinds and avoid shooting straight into a bright window.
• Avoid extreme wide angles. A very wide lens bends walls and stretches furniture, which makes a room harder to recognise and can make it look bigger than it is.
• Put the rooms in order. Upload in the order a visitor walks through the home, so the photos and their tags read as one tour.

These are practical habits, not published vendor rules; the vendors don't publish what makes their tags more accurate. Our guide on how to take real estate photos with an iPhone goes room by room.

One way to get listing photos: October

October is an iPhone app for listing photos. It guides you through each room on an iPhone 12 or newer, shows you where to stand and keeps the phone level, and you get back finished listing photos and a walkthrough video, usually within a few hours.

The finished photos are AI-altered images of your room, so label them where your MLS or state requires; the download includes the unaltered original of every photo. It costs $49 a place for photos, $99 with an HD walkthrough and $199 with a 4K walkthrough, where a place is up to 6 rooms, and your first place is free. See pricing or the demo.

Step by step

1. Remove branding

   Take logos, watermarks, names, phone numbers and websites off the MLS version of every photo.

2. Check for signs and people

   Make sure no yard sign and no person, including a reflection, is in any photo.

3. Confirm your rights

   Use only photos you took or have written permission to use, never photos copied from another listing.

4. Label altered photos

   Disclose virtual staging or AI edits the way your MLS asks, and keep the originals.

5. Check size and format

   Resize photos to your MLS's pixel and file-size limits before uploading.

6. Review the tags

   After upload, read the room labels, captions and auto-filled fields and correct anything wrong.

Questions