# AlloyDB adds database-native fusion for vector and keyword search

_Published Tuesday, October 6, 2026 at 12:07 PM EDT · AI, Infrastructure · Latest · Tier 2 — Notable_

![AlloyDB adds database-native fusion for vector and keyword search — Primary](https://storage.googleapis.com/gweb-cloudblog-publish/images/10_-_Databases.max-2600x2600.jpg)

Google says recent AlloyDB AI updates let developers combine vector and keyword search results through a single SQL function, replacing application code that normalizes scores and joins results. The function uses Reciprocal Rank Fusion, which combines results by their positions in each search ranking rather than comparing incompatible scores.

An external search connector also lets AlloyDB query Elasticsearch, OpenSearch and Solr. For internal text search, the RUM extension stores word positions in its index to support ranking and phrase matching, but requires slower index builds and more disk space. Native BM25 keyword ranking is in preview.

## Sources

- [Cloud Blog](https://cloud.google.com/blog/products/databases/simplify-ai-search-with-alloydb-hybrid-search-and-rrf/)

---
Canonical: https://techandbusiness.org/newswire/4HGGM-KbuxgBiRs7AesM2B
Published: 2026-10-06T16:07:02.845Z
Story chronology: 2026-10-06T07:00:00.000Z
Retrieved: 2026-10-06T18:14:37.243Z
Publisher: Tech & Business (techandbusiness.org)