Room images solutions
Links room photos to standard room types; each room offer appears with its own images and duplicates are merged into one image.
Hotel images collects exterior, lobby, pool, beach, restaurant, spa and common-area photos from suppliers and links them to the hotel's HM ID. Copies of the same photo that arrive from different suppliers are grouped into one image, every image is tagged with its category, and the gallery is delivered through the API and dump files.
{
"hm_id": "HM-1048213",
"images_in": 5,
"images_unique": 3,
"images": [
{
"image_id": "HM-1048213-I001",
"tags": ["pool"],
"copies": [
{ "supplier": "A", "width": 1920 },
{ "supplier": "B", "width": 1024 },
{ "supplier": "C", "width": 800 }
]
},
{
"image_id": "HM-1048213-I002",
"tags": ["beach"],
"copies": [{ "supplier": "A", "width": 1920 }]
},
{
"image_id": "HM-1048213-I003",
"tags": ["lobby"],
"copies": [{ "supplier": "C", "width": 800 }]
}
]
} Copies of the same photo that arrive in different sizes or compression merge into one image, so the gallery never shows the same pool shot three times.
Every image is tagged as exterior, lobby, pool, beach, restaurant, spa or common area; you build a tabbed gallery from these tags.
Images are linked to the matched hotel record; even if a supplier changes or a new one is added, the gallery stays with the same hotel.
Instead of comparing photos one by one, your content team receives duplicate groups and categories ready to use.
Five hotel photos from three suppliers
supplier,hotel_code,image_url,width
A,HTL-88213,https://img.supplier-a.example/88213/pool.jpg,1920
B,40917,https://cdn.supplier-b.example/40917/pool_main.jpg,1024
C,TR-ANT-551,https://media.supplier-c.example/551/IMG_2231.jpg,800
A,HTL-88213,https://img.supplier-a.example/88213/beach.jpg,1920
C,TR-ANT-551,https://media.supplier-c.example/551/IMG_2240.jpg,800 Three deduplicated, tagged images
{
"hm_id": "HM-1048213",
"images_in": 5,
"images_unique": 3,
"images": [
{
"image_id": "HM-1048213-I001",
"tags": ["pool"],
"copies": [
{ "supplier": "A", "width": 1920 },
{ "supplier": "B", "width": 1024 },
{ "supplier": "C", "width": 800 }
]
},
{
"image_id": "HM-1048213-I002",
"tags": ["beach"],
"copies": [{ "supplier": "A", "width": 1920 }]
},
{
"image_id": "HM-1048213-I003",
"tags": ["lobby"],
"copies": [{ "supplier": "C", "width": 800 }]
}
]
} Sample data; the hotel, codes and image URLs are not real. Image quality scoring and cover image selection are planned.
Hotel photos sent by suppliers are collected under the HM ID determined by hotel mapping.
A content-based fingerprint is computed for each image; the same shot arriving in a different size or compression is grouped under one image.
An image classification model assigns each image a category tag such as exterior, lobby, pool, beach, restaurant, spa or common area.
Deduplicated images, duplicate groups and tags are returned with the hotel's HM ID in the API response and in dump files.
An agency working with several suppliers receives many photos for the same hotel. Some of them are copies of the same shot in different sizes or compression. When the same pool photo appears three times in a gallery, the page looks sloppy and becomes needlessly heavy. Removing duplicates among images is the visual side of hotel-level duplicate cleanup.
For images to be linked to the right hotel, hotel mapping comes first; every photo is linked to the hotel’s HM ID. You can find the core concepts in the hotel mapping guide and the cost of duplicate records in the cost of duplicate hotels.
Each image is tagged with a category code. The codes are language-independent; in the interface you show the label in your own language.
| Category | Code | Example use |
|---|---|---|
| Exterior | exterior | “Overview” tab |
| Lobby | lobby | “Common areas” tab |
| Pool | pool | “Pool and beach” tab |
| Beach | beach | “Pool and beach” tab |
| Restaurant | restaurant | “Dining” tab |
| Spa | spa | “Spa and wellness” tab |
| Common area | common_area | “Common areas” tab |
Tags also allow a missing-image check: if a hotel’s factsheet lists a beach but no image is tagged as beach, you can raise an alert for the content team.
Photos of rooms belong not in this gallery but in the room images service, which links photos to standard room types.
| Feature | Status |
|---|---|
| Duplicate image detection | Available |
| Category tagging | Available |
| Image quality score | Planned |
| Cover image selection | Planned |
| Multilingual alt text | Planned |
This version does not score blurry or low-resolution images, does not pick a cover image for the hotel and does not write alt text for images. At this stage you choose the cover image yourself.
Images, duplicate groups and tags are returned in the API together with the hotel content response; they are also delivered in dump files for the whole inventory. The include and tag parameters in the example are an API v1 draft. For scope and pricing, you can request a quote.
Request
curl "https://api.hotelmapping.org/v1/hotels/HM-1048213/content?include=images&tag=pool" \
-H "X-API-Key: $HM_API_KEY" Response
{
"hm_id": "HM-1048213",
"images": [
{
"image_id": "HM-1048213-I001",
"tags": ["pool"],
"url": "https://img.supplier-a.example/88213/pool.jpg",
"width": 1920,
"copy_count": 3
}
]
} Code samples are illustrative; field names may change in the v1 release.
Hotel images are the hotel's general gallery: exterior, lobby, pool, beach, restaurant, spa and common areas. Room images are photos linked individually to each standard room type, and they are a separate service.
Copies that arrive in different sizes and compression are in scope today. How far cropped or watermarked copies are caught will be shared together with test results.
Today, images that come from suppliers are processed. Processing images uploaded by the agency itself is not offered yet.
In the API, tags are returned as language-independent codes (e.g. pool, beach); you display them in your own language in the interface. Multilingual alt text for images is planned.
Request a quote
Send your request in three steps; we will email the quote to your work address. The request is non-binding.