Project
Media Bank
Tourism · Photo & video
Photos and videos in one library. AI does the tagging.
A media bank for a tourism organization: a private photo and video library for the photo desk, press office, social media team and partners. We built the catalog with albums and license filters, uploads from a computer and folder import from cloud storage, automatic file processing with AI tags, a processing console and access management.
Scale and technology
Every uploaded file goes through a pipeline with no manual work: crops, content recognition and auto-tags. Photos and videos can be found by tags straight away, and the folder structure from cloud storage becomes albums.
All albums on one screen: sections on the left, filters on top, file cards with a score, license and auto-tags.
A photo opens in the viewer next to its license status, album, file name, date taken, contract and the tags assigned by AI. Below are the EXIF data: camera, aperture, shutter speed, ISO. The original downloads with one button.
In edit mode, albums are renamed, moved and deleted right in the sidebar. Orientation and license filters work inside an album.
An album is created at the top level or inside an existing section, which gives nested collections such as Seasons / Winter.
A separate section for video with themed albums: panoramas, the Underground, theatre, festivals, Reels and Shorts. Each preview shows 9:16 or 16:9 orientation and the clip length.
Files are dragged into the window, each with its own progress. Album, license and contract number are set for all files at once, and the date taken comes from EXIF. Unsupported formats are skipped, and the upload can run in the background.
Once cloud storage is connected, the media bank scans its folder structure in the background. Access is read-only: the files in storage stay untouched.
Folders are shown as tiles with file counts. Selected folders become albums or go into an existing album, and the license is set for all of them at once.
How many files were uploaded, how many were already in the media bank and how many failed after retries. New files go straight to automatic processing.
Every file goes through a pipeline: upload, crops, AI recognition and auto-tags. The table shows the status of each step and the number of attempts, and a failed step can be re-run. The system assigns tags on its own.
A list of users with access levels and sign-in counts: administrators, editors and external users such as partner agencies.
Since 2013 we have launched, sold and operated our own services. That changes how we work on yours.
We weigh every feature against cost and payback — and push back on scope that doesn’t serve the goal.
The architecture we recommend is the one running our own products under real load.
We operate our own services 24/7, so monitoring, releases and incidents are routine, not an upsell.
Pricing, payments, analytics, growth — we have been through it ourselves and speak your language.