TRONDHEIM, NORWAY -- Vespa.ai, developer of the leading platform for AI applications including Retrieval-Augmented Generation (RAG), announced support for ColPali, a new open-source retrieval model for visually rich documents such as PDFs.

ColPali enhances document retrieval by embedding entire rendered documents, including visual elements, into vector representations optimized for Large Language Models (LLMs). This reduces latency, improves accuracy, and enables more context-aware information retrieval, especially for visually rich content. By treating documents as visual entities rather than text, ColPali eliminates complex preprocessing, preserves visual context, and streamlines the RAG pipeline.

Jon Bratseth, CEO and Founder, Vespa.ai: “With ColPali’s capabilities, combined with our scalable architecture and hybrid search, Vespa.ai delivers the fastest and most accurate solution for large-scale RAG and generative AI applications. Vespa is available as a service to simplify deployment further.”