Vespa

Vespa is an AI search platform for developers building search, RAG and recommendation applications over large datasets that change continuously. It retrieves candidate results and evaluates their relevance with machine-learned models within the same platform.

• Hybrid search: Combine text, vectors and structured data to find results, with relevance models and multi-vector representations for selecting RAG context.
• Model-based ranking: Evaluate machine-learned models during queries, including computations over tensors, and distribute that work across multiple nodes.
• Recommendations and personalization: Retrieve eligible content and score it with models for recommendations, personalized results and ad targeting.
• E-commerce search: Combine search and recommendations across text, images and structured data, with structured filters to narrow results.
• Personal search: Use streaming search without building indexes when each query accesses only a small fraction of the overall dataset.

Vespa can organize and aggregate results as part of a query. Its distributed architecture supports datasets spread across multiple nodes and parallel model evaluation, making it suited to applications that need to serve queries while their underlying content keeps changing.

The platform is open source under the Apache 2.0 license and can run on your own servers. Vespa Cloud is the managed alternative, where your application runs in the cloud.