Yandex’s ‘Alice’ Helps the Visually Impaired ‘See’ Images Online
Yandex’s voice assistant, Alice, has received a major update: it can now describe images in detail, analyzing objects, their colors, and spatial arrangement. The new AI-powered feature helps blind and visually impaired users navigate online visuals — from paintings and illustrations to marketplace product photos.
Under the Hood: How Alice ‘Sees’
Alice’s neural network has reached a new level of understanding. It can now interpret what’s depicted in online images and generate detailed descriptions.
The assistant identifies what is shown in a photo or illustration and, upon request, provides additional details — such as describing the color, shape, or fit of a dress in an online store. At the core of this functionality lies a multimodal model capable of analyzing both text and images.
Unlike traditional recognition systems that focus on individual objects, Alice processes entire images while understanding relationships between elements. This context-aware approach allows the AI to produce concise yet meaningful descriptions, helping visually impaired users form an accurate mental image of what’s on screen.
Accessibility in Action
The system’s interface is designed for simplicity. Users can trigger the feature by selecting ‘Describe image’ in the context menu or by hovering over a picture using a screen reader. In Yandex Browser for Windows, iOS, and Android, Alice can both generate and voice descriptions.
The feature can be managed in the browser’s Accessibility settings. When a screen reader is active, Alice only describes images that the user focuses on. According to Yandex’s developers, the goal is to uphold accessibility principles — making the Internet an inclusive, engaging space for everyone.
The browser also includes additional features for people with visual impairments: a reading mode that removes distractions, and the ‘Smart Camera’ tool, which can recognize printed text and everyday objects. Today, Yandex Browser runs on more than 90 million devices every month.
Driving Inclusive Design Forward
The new feature enhances accessibility and user convenience, particularly for those relying on voice-based interfaces. It also demonstrates the rapid progress of Russian IT development, showcasing hybrid systems that merge computer vision with language models.
Future updates could include tighter integration with other Yandex services and improvements in object recognition — potentially expanding to automated metadata generation for images.
The Russian-developed multimodal assistant may find application in e-commerce, education, and media sectors across the CIS and Global South. To reach international markets, Yandex could release an API or SDK package, allowing third-party developers to integrate Alice’s accessibility features into their own platforms.
