Case Study: deviantART achieves AI-powered content discovery at scale with Vespa

deviantART powers AI content discovery across 1B+ documents with Vespa

deviantART, a Wix subsidiary and one of the world's largest online art communities, faced the immense challenge of scaling content discovery across its vast catalog of over 650 million artworks. With tens of thousands of new uploads daily, their platform needed to go beyond traditional search to deliver real-time, personalized results that understood artistic styles and community trends.

Vespa provided an AI-native discovery platform that integrated machine learning, custom query processing, and real-time ranking into a single high-performance serving layer. The solution enabled highly personalized discovery and domain-specific relevance. As a result, Vespa manages over 1 billion documents, serves nearly 4,000 queries per second with 12ms average latency, and ingests new content with sub-millisecond feed latency, powering a responsive experience for a global community.

deviantART