What is Vespa AI?

Vespa AI is an open-source big data serving engine built for applications that require real-time computation over large datasets, including search, recommendation, and personalisation use cases at scale. It evaluates machine learning models at query time rather than at index time, enabling results that adapt to the current query context rather than static pre-computed rankings. For teams using the Agentic Frontend Management Platform, Vespa AI provides the search and recommendation computation layer that can serve highly relevant results across large product catalogues or content libraries.

How it works

Vespa AI combines an inverted index with a document store and a tensor computation engine in a single serving layer, allowing teams to run vector similarity search, BM25 text ranking, and custom ML model scoring within a single query. Relevance logic is defined in Vespa's ranking expressions, which can blend structured filters with learned embeddings. This makes it practical to deliver Personalization at the query level, selecting results not just by keyword match but by account-level or session-level signals drawn from the B2C Growth Kit.

Vespa AI with Laioutr

Vespa AI is planned for the Laioutr App Store. Once available, Vespa AI's real-time ranking capabilities will connect to Laioutr's Agentic Frontend Management Platform, enabling search and recommendation components to serve query-time personalised results without a separate batch processing step. Teams will configure relevance models through the Laioutr integration layer and surface results through standard Personalization components, keeping the full search and recommendation pipeline within a composable architecture.

Related

Explore B2C Growth Kit.