pg_textsearch brings 4x faster BM25 search to Postgres

Native, high-performance BM25 in Postgres is a massive win for simplifying data architecture and hybrid search workflows.

• –Removing Elasticsearch from the stack dramatically reduces operational complexity and data synchronization headaches
• –The 4x performance boost via memtables addresses the historical sluggishness of native Postgres full-text search at scale
• –When paired with pgvector, this creates a formidable, unified foundation for RAG applications requiring both semantic and keyword retrieval
• –Built by Timescale, it signals a broader trend of "just use Postgres" for specialized workloads previously requiring dedicated infrastructure