Metadata updates

Vectara has enhanced its platform by introducing updateable metadata capabilities, allowing users to efficiently manage document-level metadata through standard PUT and PATCH REST semantics. This development builds on Vectara's existing features, such as hybrid search, which combines semantic search with BM25 keyword matching, and the ability to attach and filter metadata, tools that help reduce hallucinations and costs in generative AI systems. The new metadata management feature enables users to update or replace metadata without needing to delete and reindex documents, streamlining processes for use cases like price updates, security permissions, and product promotion. Currently limited to document-level metadata, Vectara plans to expand this capability to section-level metadata in the future, while also exploring additional features for metadata and document text updates. This enhancement marks a significant step in facilitating the development of GenAI applications, and Vectara encourages user feedback to further refine its offerings.