Powering Smart Photo Organization for Turkcell Lifebox

Turkcell Lifebox, Turkcell’s personal cloud storage application, needed an AI-powered visual intelligence layer to help users organize large-scale photo libraries more easily.

As users uploaded millions of photos, Lifebox aimed to improve the photo organization experience by supporting automated grouping, visual content categorization and smarter album creation. The objective was to help users rediscover their memories more easily within a large-scale consumer cloud storage environment.

Because Lifebox operates at consumer scale, the solution needed to process millions of photos per day with high accuracy, high availability and integration flexibility. It also needed to run within Turkcell’s own data center infrastructure as an on-premises deployment to support data control, privacy and operational requirements.

Nevalabs provided Uggla Visual Intelligence API as an on-premises AI software layer integrated into Turkcell Lifebox through REST APIs.

The agreement was signed on September 19, 2017, and the system went live within the following months. Uggla supported Lifebox by providing visual intelligence signals that helped group visually similar faces, classify visual content and enrich photo organization workflows.

These visual intelligence capabilities helped Lifebox support smarter album creation and memory experiences within its own application layer. Nevalabs provided the AI-based visual analysis API, while Turkcell Lifebox used these signals within its own user experience, application logic and product workflows.

Since the initial go-live, the collaboration has continued with ongoing software updates, maintenance and technical support, supporting Lifebox’s long-running visual intelligence operations at consumer cloud scale.

Uggla’s visual classification capabilities were designed to support broad photo organization needs by analyzing visual content categories and grouping them into curated structures for smart album and media organization workflows. Production usage was managed through reliability thresholds, with selected categories enabled or disabled based on accuracy, product fit and user experience considerations.

Over time, additional visual quality and face-related organization signals were added to support improved photo grouping, smart album workflows and memory generation support. Combined with Turkcell’s own application metadata, such as date and user-side photo organization logic, these signals helped support a more meaningful photo experience for Lifebox users.

The original production deployment was designed to run on Turkcell-selected on-premises server infrastructure with Intel CPU-based processing. Uggla’s CPU-optimized architecture helped deliver high-throughput visual analysis within Turkcell’s data center environment.