AI Data Platform
6.9k 2026-04-18
vespa-engine/vespa
A high-performance AI data platform for real-time search, recommendation, and machine learning inference at any scale.
Core Features
Real-time search and inference over diverse data types (vectors, tensors, text, structured data).
Scalable and highly available for large, continuously changing datasets.
Evaluates machine-learned models at serving time with low latency.
Supports complex use cases like personalization and recommendation engines.
Provides a platform for building AI-driven applications.
Detailed Introduction
Vespa is an open-source AI data platform designed for real-time search, recommendation, and personalization at any scale. It enables the selection, evaluation of machine-learned models, organization, and aggregation of data (vectors, tensors, text, structured data) with sub-100ms latency, even with continuously changing, large distributed datasets. Developed over many years and powering large internet services, Vespa provides high availability and performance for complex AI-driven applications, making it ideal for demanding online services.