← Back to blog

How to Make a Real Estate Video with AI

A 4-step AI workflow for real estate listing videos: animate the real listing photos (never invent rooms), add neighborhood b-roll, then voice-over and export. A 60-second listing video costs about 1,415 credits at 8frame's prices.

Correction, 2026-10-01: the neighborhood b-roll step gave a text-only prompt to Kling v3. On 8frame, Kling v3 needs a start image, so a text-only b-roll prompt goes to Kling 2.6 Pro (47 credits for 5 silent seconds). The walkthrough now uses it: 94 credits for that row instead of 114, and 1,415 in total instead of 1,435.

You can make a real estate listing video with AI from the photos in your MLS package: animate each real photo into a slow camera move with Seedance 2.0 or Kling v3, add neighborhood atmosphere that clearly isn't the house, and finish with a voice-over or music. The rule that matters more than any model choice: every shot of the property must come from a real photo of the property. A 60-second video with two takes per shot costs about 1,415 credits at 8frame's prices, roughly $14.

Correction, 2026-09-30: the first version of this page suggested generating exterior and interior shots of the listing from text when the photos didn't cover them (that invents parts of a real property and risks misrepresentation), suggested turning a dusk photo into daytime, described a timed run with results we can't reproduce, priced clips at $0.32 to $0.65 (57 to 108 credits), quoted videographer prices and MLS rules we can't source, and pointed to a template that doesn't exist. This version keeps the method with those fixed.

TL;DR

The accuracy rule

A listing video that shows a room, a view or a finish that isn't there is a misrepresentation, whatever tool made it. So: property shots come only from real photos of the property, animated with camera motion, not changed. Exposure fixes on a photo (recovering a blown-out window) are one thing; adding furniture, changing the season, the time of day or the view is another, and many MLS boards have rules on altered or virtually staged images. Check your MLS and local rules, and disclose anything that isn't a straight photograph.

The 4-step workflow

Step 1: Animate listing photos

Crop each listing photo to 16:9 first (Kling v3 follows the start image's shape, and one shape keeps the cut consistent), then use it as the start image.

Seedance 2.0 (4 to 15 seconds, 48 credits for 5 seconds at 480p, 108 at 720p) or Kling v3 Standard (3 to 15 seconds, 57 for 5 silent seconds).

Prompt structure: [Room as it appears in the photo]. Slow [direction] camera move, holds composition. [Lighting as in the photo]. No warping on walls or furniture. Static furniture.

Primary bedroom:

A spacious primary bedroom with light hardwood floors, white walls, and a king bed centered under a window. Slow dolly-in toward the bed from the doorway. Warm afternoon light through sheer curtains, soft shadows. Static furniture, no motion artifacts. No wall distortion on edges.

Kitchen:

A bright kitchen with white shaker cabinets, marble countertops, and stainless steel appliances. Camera pans left to right across the island, ending on the window above the sink. Natural daylight. Countertops stay flat throughout the pan.

What to check: baseboards and wall edges that warp, countertops that bend during a pan, reflections that shimmer on appliances, and anything the model added that isn't in the photo. Run 2 takes per room and keep the cleaner one.

Step 2: Exterior and wide shots, from photos only

The exterior approach and the living room hero angle are the most valuable shots; make them from the real exterior and living room photos, with a slow push or a gentle descent.

Two-story craftsman house, white trim, covered front porch, set back from street behind a mature oak tree. Camera slowly moves toward the front door. Late afternoon light, long lawn shadows. House stays geometrically stable throughout.

What to check: rooflines and window grids that stay straight, and a tree that doesn't flicker. If you don't have a photo of an angle, you don't have that shot; don't generate it.

Step 3: Neighborhood b-roll

Atmosphere that supports the story without claiming to be the property or its street: a generic morning street, a park, a coffee shop interior.

A quiet suburban street in early morning. A couple walks a golden retriever along the sidewalk, dappled light through mature street trees. Slow pan following the couple from behind. Warm golden morning light, long shadows.

Keep it obviously generic, and keep it short. This prompt is text only, and on 8frame Kling v3 needs a start image, so run it on Kling 2.6 Pro: 47 credits for 5 silent seconds. Fair-housing rules apply to how listings depict people and neighborhoods; keep b-roll about place and light, not about who should live there.

Step 4: Voice-over or music, and export

A voice-over suits listing sites where buyers are comparing; music suits social. ElevenLabs text to speech costs 13 credits per started 1,000 characters (a 60-second script fits in one), ElevenLabs Music 81 per started minute.

A 60-second cut: exterior approach (0 to 4s), living room (4 to 9s), kitchen (9 to 15s), primary bedroom (15 to 21s), other rooms (21 to 38s), neighborhood (38 to 48s), exterior hold with agent contact (48 to 60s).

Export 16:9 at the clips' resolution for listing sites and email, and a 9:16 version for Reels and TikTok. Room shots usually keep the subject in the center third, so a centered crop often works; check each shot.

Routing by property type

Luxury: animate every interior photo, 2 to 3 takes per room, and consider a real drone pass for aerials. AI covers the interiors from photos; it can't fly over the real lot.

Mid-market and starter: the workflow above, 2 takes per shot. This is where AI makes video affordable on every listing.

Commercial (office, retail, industrial): wide, high-ceiling spaces from photos with slow pushes; neighborhood b-roll shifts to commute and street context.

Walkthrough: a 3-bed house, 60-second listing video

Inputs: the MLS photos covering exterior front and back, living room, kitchen, bedrooms, primary bath, backyard.

Shot (all from real photos except the last video row) Model Takes Credits
Exterior push-in Kling v3, 5 s silent 2 114
Living room descent Kling v3, 5 s silent 2 114
Kitchen pan Seedance 2.0, 5 s at 720p 2 216
Primary bedroom dolly Seedance 2.0, 5 s at 720p 2 216
Primary bath push Seedance 2.0, 5 s at 720p 2 216
Bedroom 2 Seedance 2.0, 5 s at 720p 2 216
Backyard pan Kling v3, 5 s silent 2 114
Neighborhood b-roll (generic, not the property) Kling 2.6 Pro from text, 5 s silent 2 94
Exposure-fixed exterior still Nano Banana Pro 1 21
Voice-over ElevenLabs text to speech 1 13
Music bed ElevenLabs Music 1 81
Total 1,415, about $14

At 480p on Seedance the four interior shots cost 384 instead of 864, 955 in total.

Pitfalls

Window blowout. A blown-out window in the photo animates as a white void. Recover the exposure on the still first (Nano Banana Pro, "recover window exposure, change nothing else", 21 credits) and check the result against the original.

Room scale drift. Tight dolly-ins in small rooms can make the space look like a dollhouse by the end. Add "furniture proportions consistent throughout, room scale stable" and run takes.

Mirrors. Mirrors animate unpredictably: doubled camera moves or frozen reflections. Frame mirrors at the edge, or pick a different photo.

Invented details. Watch every take for things that aren't in the photo: an extra window, a different floor, a longer room. Discard those takes.

FAQ

Are AI listing videos allowed by the MLS?

Rules vary by MLS and jurisdiction, and some boards have rules on altered or virtually staged images, so check yours. The safe principle: the video must show the actual property. Animating real listing photos with camera motion keeps the rooms real; generating rooms or views does not.

Can I use my existing MLS listing photos?

Yes, they're the right input. Crop them to 16:9 and use each as the start image. An average real photo beats any text description of a room.

What aspect ratio should I use?

16:9 for listing sites, MLS embeds and email; 9:16 for Reels, TikTok and Stories, exported from the same edit with a centered crop you check shot by shot.

Sources


Animate what's real, and let the neighborhood carry the atmosphere. For the wider picture, see the AI video for ecommerce complete 2026 guide. The 8frame canvas is free and unlimited; generation is paid from $19/month.

Written with AI assistance. Corrections are noted on the page. Editorial policy

Related articles

use caseHow to Make a Restaurant Video with AIuse caseHow to Make an ASMR Product Video with AIuse caseHow to Make a Shopify Product Video with AI

Make it
move.

Stay in the loop

Be the first to hear about our launch and get product updates