Google Cloud Brings Native BM25 Full-Text Search to AlloyDB and Cloud SQL via Tiger Data’s pg_textsearch
Native BM25 ranking in managed databases validates pg_textsearch as an industry standard for PostgreSQL full-text search
Press Release Disclaimer: This is a press release distributed through the XPR Media network. It has not been independently verified by our newsroom.
NEW YORK, Sept. 21, 2026 (GLOBE NEWSWIRE) — Tiger Data, the creators of TimescaleDB, today announced that its open-source BM25 search engine, pg_textsearch, is now supported natively in Google Cloud’s AlloyDB and Cloud SQL for PostgreSQL databases. The integration makes industry-standard full-text ranking available natively in two of the world’s most widely used managed PostgreSQL services, eliminating the need for developers to maintain separate search infrastructure.
The pg_textsearch extension has become a top choice for full-text search in PostgreSQL. Unlike earlier Postgres search methods that degrade with scale, BM25 scoring applies inverse document frequency, term frequency saturation, and length normalization, three mathematical properties that make rankings more accurate. This extension supports PostgreSQL 17 and 18 and stores indexes in standard PostgreSQL pages, which means full integration with WAL logging, replication, VACUUM, and pg_dump. No external storage, special backup procedures, or operational overhead are needed.
Google Cloud has integrated pg_textsearch into AlloyDB and Cloud SQL, which now allows developers to use a single, unified search backend rather than sidecar systems. pg_textsearch achieves this in pure PostgreSQL.
“Developers have spent years bolting on or maintaining separate search clusters just to get relevance ranking that works,” said Mike Freedman, CTO and co-founder of Tiger Data. “With pg_textsearch now available across managed services like AlloyDB and Cloud SQL it sends a clear message: this is the search architecture the industry is converging on. It’s a validation that native, database-integrated full-text search, built in C on PostgreSQL’s storage layer outperforms wrapping external libraries and dealing with the operational complexity that comes with them.”
AlloyDB and Cloud SQL customers can now harness native BM25 ranking alongside pgvector for seamless hybrid search. They can also gain access to ScaNN vector indexing, which accelerates vector search speeds by up to six times compared to standard HNSW. Both AlloyDB and Cloud SQL can merge keyword and semantic results in a single query using Reciprocal Rank Fusion, eliminating the need for separate backends and sync pipelines.
pg_textsearch is now available in public preview on both AlloyDB and Cloud SQL, with general availability expected later this year.
Learn more:
- Google Cloud blog post announcing the inclusion of pg_textsearch in AlloyDB and Cloud SQL: https://cloud.google.com/blog/products/databases/native-bm25-search-in-alloydb-and-cloud-sql
- AlloyDB BM25 documentation: https://docs.cloud.google.com/alloydb/docs/ai/create-bm25-index
- Cloud SQL BM25 documentation: https://docs.cloud.google.com/sql/docs/postgres/pg-textsearch
- pg_textsearch on GitHub: github.com/timescale/pg_textsearch
- Tiger Data pg_textsearch guide: tigerdata.com/docs/learn/search/using-pg-textsearch
About Tiger Data
Tiger Data, creators of TimescaleDB, helps teams build applications that combine relational and time-series data in Postgres. Built for demanding sensor and machine data workloads, TimescaleDB combines PostgreSQL and transactional guarantees with capabilities for high-volume ingestion, time-series queries and long retention. Organizations can keep machine data, business records and application transactions together, with deployment options spanning the fully managed Tiger Cloud service, TimescaleDB Enterprise for on-premises and edge environments, and self-managed TimescaleDB community edition. Learn more at tigerdata.com.
Media Contact:
Christine Penwell
Offleash PR for Tiger Data
