Real estate listing creation and virtual tour generation

Introduction

Creating real estate listings is repetitive work. A property manager receives photos, writes a description, uploads it to a listing platform, then has to generate a virtual tour separately. Each step requires manual context switching, and inconsistencies creep in between description and tour footage. For agencies managing dozens of properties monthly, this process burns hours that could go elsewhere.

The workflow we're automating here takes raw property data and images, generates professional listing copy through AI, creates a video virtual tour, and organises everything for publishing. No human intervention between upload and final output. This is where combining AI-Boost (for data enrichment), Hour One (for video generation), and WeryAI (for structured listing creation) becomes practical.

The orchestration layer ties these three tools together so that when you drop property details into a spreadsheet or webhook, everything else happens automatically. We'll walk through the technical setup using either Zapier, n8n, or Claude Code depending on your preference for simplicity versus control.

The Automated Workflow

Overview of the Process

The workflow operates in this sequence: property metadata arrives via webhook or spreadsheet import; AI-Boost enriches the raw data and generates compelling descriptions; Hour One converts those descriptions into a video presentation with generated speech and visuals; WeryAI structures everything into a formatted listing; the final output lands in your database or listing portal.

The entire chain completes within 10 to 15 minutes depending on video generation time. No manual steps, no copy-pasting between tools.

Choosing Your Orchestration Tool

Zapier is best if you want speed and don't need complex branching logic. It has pre-built connectors for most platforms and handles authentication painlessly.

n8n suits teams wanting self-hosting and full visibility into the data at each step. The JSON editor makes it straightforward to transform data between tools.

Make (Integromat) sits between the two; it's more visual than n8n but more flexible than Zapier.

Claude Code is an option if you're building this once as a script rather than a recurring automation; it gives you maximum control but requires manual triggering.

For this workflow, we'll show the n8n approach since it gives you the best balance of transparency and doesn't lock you into monthly connector costs.

Step 1:

Setting Up the Webhook Entry Point

Your automation starts when property data arrives. This could be a form submission, an API call, or a scheduled import from a spreadsheet.

Copy
POST /webhook/property-listing
Content-Type: application/json

{
  "property_id": "PROP-2024-001",
  "address": "42 Maple Lane, Bristol",
  "bedrooms": 3,
  "bathrooms": 2,
  "square_feet": 1850,
  "price": 425000,
  "key_features": ["Period property", "Garden", "Off-street parking"],
  "image_urls": ["https://storage.example.com/prop1.jpg", "https://storage.example.com/prop2.jpg"]
}

In n8n, create an incoming webhook node and set it to POST. This becomes the trigger point for the entire workflow. Every piece of property data flowing through here will kick off the automation.

Step 2:

AI-Boost for Description Enrichment

AI-Boost takes structured property data and generates marketing-ready descriptions. The API expects property attributes and returns polished copy suitable for listing platforms.

The endpoint looks like this:

Copy
POST https://api.ai-boost.io/v1/generate/real-estate-description
Headers:
  Authorization: Bearer YOUR_API_KEY
  Content-Type: application/json

{
  "property_data": {
    "address": "42 Maple Lane, Bristol",
    "property_type": "Victorian terraced house",
    "bedrooms": 3,
    "bathrooms": 2,
    "square_feet": 1850,
    "price": 425000,
    "year_built": 1890,
    "key_features": ["Period property", "Original cornicing", "Garden", "Off-street parking"],
    "condition": "Well-maintained"
  },
  "style": "professional",
  "word_count": 200,
  "tone": "engaging"
}

AI-Boost returns something like:

Copy{
  "description": "Step into Victorian charm with this beautifully preserved three-bedroom terraced house in desirable Bristol. Original cornicing and period details create character throughout, whilst modern amenities ensure comfortable living. The mature garden offers peaceful outdoor space, and dedicated parking provides practical convenience. An exceptional opportunity for buyers seeking authentic period properties with contemporary comfort.",
  "highlights": ["Period features", "Garden access", "Parking included"],
  "seo_keywords": ["Victorian property", "Bristol terraced house", "period home"]
}

In n8n, add an HTTP request node configured with the AI-Boost endpoint. Map your incoming webhook data to the required fields. The response gives you the description and SEO keywords to use later.

Step 3:

Hour One for Video Generation

Hour One creates video presentations from text scripts and images. It works by generating a talking-head video or presentation-style tour narrated by a digital avatar or text-to-speech voice reading your property description.

The API for creating a video project:

Copy
POST https://api.hourone.com/v1/projects
Headers:
  Authorization: Bearer YOUR_HOUR_ONE_API_KEY
  Content-Type: application/json

{
  "name": "Virtual Tour - 42 Maple Lane",
  "script": "Step into Victorian charm with this beautifully preserved three-bedroom terraced house...",
  "avatar": "maya",
  "voice": "en-GB-female",
  "background_images": [
    "https://storage.example.com/prop1.jpg",
    "https://storage.example.com/prop2.jpg",
    "https://storage.example.com/prop3.jpg"
  ],
  "duration_per_slide": 4,
  "video_format": "1080p"
}

Hour One responds with:

Copy{
  "project_id": "proj_abc123xyz",
  "status": "processing",
  "video_url": null,
  "estimated_completion": "2024-01-15T14:30:00Z"
}

Videos typically render within 5 to 10 minutes. You'll need to poll the status endpoint or wait for a webhook callback to know when the video is ready.

In n8n, after getting the Hour One project ID, add a "Wait" node or schedule a follow-up HTTP request to check the status. Once status equals "completed", extract the video_url for the next step.

Copy
GET https://api.hourone.com/v1/projects/{project_id}
Headers:
  Authorization: Bearer YOUR_HOUR_ONE_API_KEY

Step 4:

WeryAI for Structured Listing Output

WeryAI formats everything into a structured, platform-ready listing. It takes your description, video URL, images, and property metadata, then outputs a JSON object or HTML snippet compatible with major listing portals.

Copy
POST https://api.weryai.io/v1/listings/create
Headers:
  Authorization: Bearer YOUR_WERYAI_API_KEY
  Content-Type: application/json

{
  "property_id": "PROP-2024-001",
  "address": "42 Maple Lane, Bristol",
  "price": 425000,
  "description": "Step into Victorian charm...",
  "bedrooms": 3,
  "bathrooms": 2,
  "square_feet": 1850,
  "listing_type": "for_sale",
  "images": [
    "https://storage.example.com/prop1.jpg",
    "https://storage.example.com/prop2.jpg"
  ],
  "virtual_tour_video": "https://output.hourone.com/videos/proj_abc123xyz.mp4",
  "seo_keywords": ["Victorian property", "Bristol terraced house"],
  "agent_name": "John Smith",
  "agent_phone": "+44 117 XXX XXXX"
}

WeryAI returns a completed listing object:

Copy{
  "listing_id": "LIST-2024-001",
  "status": "ready_to_publish",
  "listing_url": "https://listings.example.com/properties/list-2024-001",
  "html_embed": "<div class='listing'>...</div>",
  "structured_data": {
    "@context": "https://schema.org",
    "@type": "RealEstateAgent",
    "name": "42 Maple Lane Property Listing"
  }
}

This becomes your final output. Store the listing_url and listing_id in your database.

Complete n8n Workflow Configuration

Here's how your n8n workflow nodes should connect:

1. Webhook trigger node (incoming property data)
2. AI-Boost HTTP request (enrich and generate description)
3. Hour One HTTP request (create video project)
4. Wait node (5 minute delay for video processing)
5. Hour One status check (HTTP request to fetch completed video)
6. Condition node (check if video status is "completed")
7. WeryAI HTTP request (create final listing)
8. Database update or Slack notification (log the result)

The condition node ensures you don't send incomplete videos to WeryAI. If the video isn't ready after 5 minutes, you can retry or send an alert to your team.

Here's a sample n8n expression for the WeryAI step that pulls data from all previous nodes:

Copy{
  "property_id": "{{ $node.Webhook.json.property_id }}",
  "address": "{{ $node.Webhook.json.address }}",
  "price": {{ $node.Webhook.json.price }},
  "description": "{{ $node.AiBoost.json.description }}",
  "bedrooms": {{ $node.Webhook.json.bedrooms }},
  "bathrooms": {{ $node.Webhook.json.bathrooms }},
  "square_feet": {{ $node.Webhook.json.square_feet }},
  "images": {{ JSON.stringify($node.Webhook.json.image_urls) }},
  "virtual_tour_video": "{{ $node.HourOneStatus.json.video_url }}",
  "seo_keywords": {{ JSON.stringify($node.AiBoost.json.seo_keywords) }}
}

The Manual Alternative

If you're not comfortable setting up webhooks and orchestration, you can still capture the time savings with a semi-automated approach.

Create a Google Form that collects property details (address, bedrooms, bathrooms, image URLs). Use Zapier's free tier to move responses into a spreadsheet. Then manually run each tool in sequence: paste the spreadsheet data into AI-Boost's web interface to get the description, copy that description into Hour One's studio to generate the video, and finally input everything into WeryAI's dashboard to format the listing.

This saves you the description-writing work and video editing but still requires some manual copy-pasting. It's simpler to set up but takes longer per property; suitable if you're handling fewer than five new properties monthly.

The fully automated version only makes sense if you're processing listings regularly enough that the orchestration setup pays for itself within a few weeks.

Pro Tips

Rate Limiting and Throttling: Hour One processes videos sequentially. If you're submitting 10 property videos simultaneously, they'll queue and complete more slowly. Stagger your incoming webhook submissions using a delay node in n8n, or use Zapier's built-in throttling to process one property every 15 minutes. Check each API's documentation for rate limits; AI-Boost typically allows 100 requests per minute on paid plans.

Error Handling and Retries: Video generation occasionally fails due to image encoding issues or service unavailability. Add retry logic to your n8n workflow so failed Hour One requests automatically reattempt after 2 minutes, up to three times. For critical steps like WeryAI, log any failures to a separate "failed listings" channel in Slack so your team knows to investigate.

Testing with Sample Data: Before running the automation on live property data, test the entire workflow with dummy information. Create a test property with placeholder images and run it through end-to-end. This catches configuration errors and API authentication issues before they affect real listings.

Image Quality and Size: Hour One video generation works best with images between 1MB and 5MB; larger files slow down processing without visible quality gain. Consider adding an image optimisation step using a free tool like TinyPNG's API if your source images are consistently oversized. Aspect ratio matters too; 16:9 works smoothly with Hour One's default video output.

Cost Optimisation: WeryAI charges per listing created. If you're generating listings for multiple platforms (Rightmove, Zoopla, etc.), structure your workflow so it creates one master listing in WeryAI, then distributes it downstream rather than creating duplicates. Similarly, if you have properties with multiple image sets, consider batching video generation so one script with multiple images creates a richer tour than several short ones.

Cost Breakdown

Total for the full automated stack using n8n self-hosted: approximately £227 monthly. A small agency processing 20 new properties monthly saves roughly 30 hours on description writing and tour generation alone, paying for the software within the first month.

If you're using Zapier instead of n8n, add £51 and remove the n8n cost; the total becomes £278. The trade-off is convenience versus cost; Zapier requires zero server maintenance but you'll hit task limits faster as you scale.