Vespa
An AI search and serving platform for organizing and querying vectors, tensors, text, and structured data at scale, with sub-100-millisecond responses over large, continuously changing datasets.
Vespa is a platform for search, recommendation, and personalization use cases that need to select a subset of data from a large, continuously changing corpus, evaluate machine-learned models over it, and aggregate the results, typically in under 100 milliseconds. It is designed to distribute this work across multiple nodes and evaluate it in parallel.
Applications can add their own Java components, and a new release is built from the project's master branch most weekday mornings. It can be deployed to a managed cloud service or run as a self-hosted instance.
It is written in Java and released under the Apache-2.0 license.