Vespa Newsletter, June 2025

The Vespa Newsletter for June 2025 highlights significant updates and new features aimed at enhancing AI retrieval and search capabilities. Key developments include the introduction of layered ranking for RAG applications, allowing more refined document and content chunk scoring, and new tensor functions for improved ranking processes. Generative AI retrieval is bolstered with features like integrated chunking and facet filtering, while the Pyvespa Match Evaluator enhances recall evaluation for queries. The newsletter also announces the general availability of Azure zones on Vespa Cloud, alongside AWS and GCP, and introduces new query API settings for greater flexibility. Furthermore, Vespa continues to support advanced linguistic processing and has integrated language detection using Apache OpenNLP to improve query language identification. These updates, coupled with improved tensor operations and query performance optimizations, position Vespa as a robust platform for real-time AI workloads, as recognized in a GigaOm report.