Reel Houses

The Pipeline

6 photos.
6 steps.
One cinematic tour.

Every video follows the same automated pipeline. No manual editing. No desktop software. Just APIs, prompts, and a Mac.

6
Photos in
6
Steps
<1hr
Total time
0
Humans editing
01

Select Hero Images

Pick 6 best photos from the MLS listing. The system works with any standard listing photography — no special shots required. Exterior, living room, kitchen, bedrooms, and a feature shot (pool, view, fireplace).

Input: 6 JPGs
02

Image Prep

Photos are cropped to 16:9 cinematic aspect ratio (1536×864). MLS watermarks are automatically removed via intelligent cropping. Colors are normalized for consistent tone across all 6 frames.

Automatic
03

AI Animation

Each photo is submitted to Google's Veo 3.1 video generation model with a custom cinematic prompt. The AI animates each still into a smooth, realistic video clip with specific camera movement:

  • ExteriorsSlow dolly forward, evening light shifts
  • Living roomsLateral pan, firelight flickers
  • KitchensMicro dolly in, light across countertops
  • BedroomsSubtle zoom revealing full room
  • OutdoorPanoramic pan across landscape

Each clip runs 5–8 seconds. Camera movement is anchored to real objects in the photo — the stone mantle, the granite countertop, the timber beams — preventing AI drift or hallucinated objects.

Google Veo 3.1
04

Script, Voice & Music

A property-specific voiceover script is generated (50–80 words, 30–50 seconds). Scripts follow a proven structure:

  • OpenLocation + neighborhood context
  • MiddleRoom-by-room, synced to visuals
  • Features3–4 key highlights
  • CloseSpecs + call to action

The script is voiced by ElevenLabs AI (multiple voice options). Background music is generated to match the property aesthetic — warm acoustic for mountain homes, ambient electronic for urban condos, airy percussion for beachfront.

ElevenLabs
05

Automated Assembly

A single command assembles the final video. All clips normalized to 1920×1080, 30fps, H.264. Concatenated in narrative sequence — exterior → interior rooms → feature → closing. Voiceover mixed at full volume. Music at −12dB with fade in/out.

FFmpeg
06

QA & Export

Final video reviewed for AI artifacts, timing alignment, and audio balance. Exported as broadcast-ready 1080p MP4. Upload directly to MLS, Zillow, YouTube, Instagram, TikTok — no conversion needed.

Output: 1080p MP4

The secret sauce

Cinematic prompt
engineering

The difference between a slideshow and cinema is in the prompts. Each one is crafted with:

  • Camera movementSpecific: "slow micro dolly forward"
  • Spatial anchorsTied to real objects in the photo
  • Material + light"Cedar catches warm light"
  • Duration control8s open/close, 5s interiors
Fireplace room

Tech stack

Production-grade tools

The same AI models powering Google and ElevenLabs production workflows.

Google Veo 3.1

Video generation — animates still photos into cinematic clips with realistic camera movement.

ElevenLabs

Voice + music generation — professional AI voiceover and property-matched background scores.

FFmpeg

Assembly pipeline — clip normalization, concatenation, multi-track audio mixing, broadcast encoding.

Comparison

Traditional production
vs. Reel Houses

Traditional Video Reel Houses
Cost per video $500 – $2,000+ $49
Turnaround 3–7 days < 1 hour
Crew required Videographer + editor None
Scheduling Coordinate staging, weather, time Works instantly
Voiceover $150–$500 extra Included
Licensed music $50–$200 extra Included
Reshoots Reschedule entire shoot Re-run with new photos
Scalability 1 video per shoot Dozens per day

Ready to see it
in action?

Send us 6 photos. We'll send back a cinematic tour.

Get started →