Automated real estate listing pipeline from inspection photos to virtual tour

An estate agent receives 200 property photos from a morning shoot. By afternoon, they're still editing the first 30, writing descriptions that sound like they were generated by someone with English as a third language, and wondering whether they'll have time to create a virtual tour before the evening deadline. Meanwhile, their competitor uploaded five completed listings before lunch. The gap between capturing property data and publishing it live is where estate agents lose time and money. Each step demands human attention: photo retouching, background adjustments, copywriting that sells, and video assembly. What if you could feed raw inspection photos into a system that handles everything automatically, then delivers polished listings ready to publish? This workflow combines five tools into a single pipeline that transforms unedited photos into complete property listings with video tours. The orchestration layer ensures zero manual handoff between stages, from image upload to final video generation. For more on this, see Image Editing and Generation for E-Commerce: Which Free T.... For more on this, see Real estate virtual tour generation from inspection photos. For more on this, see Real estate listing automation from property inspection r....

The Automated Workflow

The pipeline runs on n8n, which offers the flexibility to handle conditional logic (some properties need more photos than others) and built-in error recovery. Here's how data flows through the system.

Step 1: Photo Upload and Triggering

Create a webhook in n8n that listens for property photo uploads. Use an AWS S3 integration or a simpler approach with Make's file upload trigger. When photos arrive, the workflow extracts metadata: property address, room type, and image count.

CopyPOST https://n8n.yourinstance.com/webhook/property-photos
Content-Type: application/json { "property_id": "PROP_12847", "address": "142 Elm Street, Bristol BS1 2AA", "photos": [ "photo_1_raw.jpg", "photo_2_raw.jpg", "photo_3_raw.jpg" ], "rooms": ["living_room", "kitchen", "bedroom_1", "bathroom"]
}

Step 2: Image Enhancement with Pixelcut AI and AI Boost

The first processing stage calls Pixelcut AI to automatically enhance each raw photo. Pixelcut removes unwanted background clutter and improves colour balance without requiring you to specify what needs fixing. Configure the Pixelcut API call in n8n like this:

CopyPOST https://api.pixelcut.ai/v1/image/enhance
Authorization: Bearer YOUR_PIXELCUT_API_KEY
Content-Type: application/json { "image_url": "https://s3.amazonaws.com/property-photos/photo_1_raw.jpg", "enhancement_type": "auto", "background_removal": true, "color_correction": true
}

After Pixelcut processes the image, pipe it to AI Boost for selective upscaling and background replacement. If a room's background looks dated (outdated furniture visible through a window), AI Boost can generate a more appealing exterior view without looking artificial.

CopyPOST https://api.aiboost.com/v2/image/upscale
Authorization: Bearer YOUR_AIBOOST_API_KEY
Content-Type: application/json { "image_url": "https://pixelcut-output.example.com/enhanced_photo_1.jpg", "scale_factor": 1.5, "background_prompt": "modern suburban street, afternoon sunlight", "preserve_room_staging": true
}

Step 3: Property Description Generation

Pass the property metadata (address, room types, approximate square footage from your records) to Copy.ai. Use a custom template that generates descriptions in your agency's voice, not generic template language.

CopyPOST https://api.copy.ai/v1/content/generate
Authorization: Bearer YOUR_COPYAI_API_KEY
Content-Type: application/json { "template_id": "estate_agent_property_description", "variables": { "address": "142 Elm Street, Bristol BS1 2AA", "rooms": ["spacious living room", "modern kitchen", "master bedroom", "family bathroom"], "property_type": "semi-detached", "key_features": ["period features", "south-facing garden", "off-street parking"], "tone": "friendly_professional", "length": "medium" }
}

Copy.ai returns polished copy that you can use directly or refine in a manual review step (optional). Store the output in your property database.

Step 4: Video Tour Generation

Hour One transforms the property description and a sequence of photos into a video tour with a virtual presenter. This is where the workflow gets interesting: instead of hiring someone to film a walkthrough, the system generates a professional-looking video automatically. First, prepare a script segment from the Copy.ai output. Then send it to Hour One:

CopyPOST https://api.hourone.com/v1/videos/create
Authorization: Bearer YOUR_HOURONE_API_KEY
Content-Type: application/json { "presenter_id": "host_professional_female_2", "script": "Welcome to 142 Elm Street. This charming semi-detached home features period character with modern comfort. The spacious living room, visible on your left, catches afternoon sun through bay windows. The fully fitted kitchen combines traditional styling with contemporary appliances.", "background_images": [ "https://aiboost-output.example.com/living_room_enhanced.jpg", "https://aiboost-output.example.com/kitchen_enhanced.jpg", "https://aiboost-output.example.com/bedroom_enhanced.jpg", "https://aiboost-output.example.com/bathroom_enhanced.jpg" ], "video_style": "professional_tour", "duration_seconds": 90, "output_format": "mp4"
}

Hour One returns a video URL. Store this in your listing management system and trigger a notification when it's ready.

Step 5: Assembly and Publishing

The final n8n step collects all outputs (enhanced photos, description, video URL) and formats them for your property portal API or CMS. If you use Rightmove, Zoopla, or similar platforms, create a custom integration that publishes directly:

CopyPOST https://api.yourlisting-portal.com/v1/properties/create
Authorization: Bearer YOUR_PORTAL_API_KEY
Content-Type: application/json { "property_id": "PROP_12847", "address": "142 Elm Street, Bristol BS1 2AA", "description": "[Generated by Copy.ai]", "images": ["[Pixelcut + AI Boost enhanced URLs]"], "virtual_tour_video": "https://hourone-output.example.com/tour_12847.mp4", "listing_status": "live", "published_at": "2026-03-15T14:22:00Z"
}

Set up error handling at each stage. If Pixelcut fails on an image, log it and send you a Slack notification rather than halting the entire workflow. N8n's conditional nodes handle this cleanly:

CopyIF pixelcut_response.status == "error" THEN send_slack_notification("Photo enhancement failed", error_details) AND retry_with_aiboost_only
ELSE continue_to_next_step

The Manual Alternative

If you prefer human oversight at specific points, add approval gates into the n8n workflow. After Copy.ai generates descriptions, pause the workflow and send the estate agent a Slack message with the copy. They click "approve" or "edit", and the workflow continues or loops back to regeneration. This slows things down but maintains quality control. Many agencies run the first 48 hours in manual-review mode, then switch to fully automated once they're confident in the system's output.

Pro Tips

Rate Limiting and Batching.

Pixelcut and AI Boost enforce rate limits (typically 30 calls per minute on standard plans).

If you're processing 50 properties, space the API calls across 2-3 minutes using n8n's delay nodes. Group photos by property to keep related data together in your logs.

Video Length and Presenter Consistency.

Hour One's presenter should stay consistent across all your listings, building brand recognition. Choose one presenter style and stick with it. Keep videos between 60-90 seconds; longer tours bore viewers, shorter ones feel rushed.

Cost Per Listing.

A typical 4-5 room property workflow costs roughly £2-3 in API fees (Pixelcut + AI Boost + Copy.ai + Hour One). At scale, this is significantly cheaper than hiring someone for photo editing and video assembly.

Fallback Images.

If background generation fails on a specific photo, the workflow automatically skips that image and moves forward with the next one. You can fill gaps later or use the original unenhanced version as a fallback.

Testing and Iteration.

Run the workflow on five test properties first. Check that descriptions match your tone, photos look natural (not over-processed), and videos are watchable. Adjust Copy.ai prompts and Hour One presenter choice based on real results, not assumptions.

Cost Breakdown