Search & retrieval
Search, recommendations and retrieval-augmented generation on Vespa, using your own data.
Talk about search
Search, recommendations and RAG
Help people find the right products and documents, and discover related content through recommendations. We build these experiences on Vespa, combining keyword and vector retrieval with ranking tuned to your content.
For retrieval-augmented generation (RAG), we connect the retrieved information to answer generation, with citations back to the sources. Data pipelines keep indexed content up to date as your information changes.
Considering a move from Elasticsearch or OpenSearch? Explore migration to Vespa.
Try a ranking example
Choose an example, then switch between text relevance and availability-first ranking. Watch how the order changes.
Query words in product names are prioritised.
Sample results
Showing 3 results for “everyday backpack”. Text relevance ranking. First result: Everyday backpack.
Everyday backpack
€75
Out of stockTrail backpack
€80
In stockCity commuter pack
€65
In stock
Clevora, our personalized search accelerator
Built on Vespa, Clevora brings together search and personalized recommendations, using interaction signals to shape discovery. It gives us a starting point for building search experiences around your content and users.
During discovery, we assess whether Clevora fits your project.
Talk about ClevoraEvaluation and production
How we workWe agree what useful results look like for your application, then evaluate retrieval and ranking against representative queries and your data throughout delivery. We also test response times and operating costs, and build in monitoring for production.
What should your users be finding?
Tell us about your data and what a useful search result or answer would look like.