Vespa is an open-source platform for search and recommendation that stores text, structured and vector data and ranks results with machine-learned models as each query runs.
In plain words
Many search setups glue together a keyword engine, a vector database and a separate model server. Vespa does all of this in one system. You describe your documents in a schema, including how to index them and how to rank them. At query time Vespa selects matching documents across many machines, can score them with a machine-learned model, for example one in ONNX format, and groups the results. It is licensed under Apache 2.0 and also offered as a cloud service.
A quick example
Imagine a news app that must show each reader the most relevant articles. Vespa stores every article's text, date and embedding. When a reader opens the app, a query picks candidate articles by keywords and vector similarity, then a ranking model scores each one using those signals before the top results are returned.
Where you'll meet it
- Large-scale search and recommendation services.
- Hybrid text and vector search with custom ranking.
Related
Sources
- DocsOverview, Vespa documentation · read 29 Sept 2026
- Docsvespa-engine/vespa, Vespa (GitHub) · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.