Vespa
An open-source AI search platform for self-hosted retrieval and recommendations, with hybrid search, model inference and an Apache 2.0 license.
Vespa is a self-hosted AI search platform for developers building search, RAG, and recommendation systems over large, changing datasets. It combines retrieval with machine-learned ranking, so an application can find candidate results and evaluate their relevance in the same platform. The code is open source under Apache 2.0. You can run it on your own servers or use the managed Vespa Cloud service, where applications run in the cloud.
Vespa searches text, vectors and structured data together. For RAG applications, it supports hybrid search, relevance models and multi-vector representations, giving developers ways to select context beyond vector similarity alone. It also works with tensors and evaluates machine-learned models during queries.
Its distributed architecture spreads data and model evaluation across multiple nodes. It's built for workloads that need fast responses while their underlying content changes continuously, with support for organizing and aggregating results as part of serving a query. That makes it relevant to teams whose search and ranking requirements extend beyond storing embeddings.
Recommendation and personalization applications can retrieve eligible content and score it with models. E-commerce applications can combine search and recommendations with structured filters across text and image content. For personal or private search, Vespa has a streaming search mode that avoids building indexes when each query accesses only a small portion of the total dataset.
