ScyllaDB X Cloud Introduces High-Performance Full-Text Search

New Full-Text Search adds BM25-ranked keyword retrieval to ScyllaDB, complementing Vector Search for fast lexical and semantic search over real-time application data

SUNNYVALE, Calif., Oct. 06, 2026 (GLOBE NEWSWIRE) -- ScyllaDB announced Full-Text Search as part of its X Cloud release, extending ScyllaDB’s real-time data platform with high-performance keyword search.

Semantic search is essential for modern AI applications, but exact words still matter. Product names, error codes, identifiers, technical terms, and phrases are often better served by lexical search.

With ScyllaDB 2026.3, applications can now use BM25-ranked Full-Text Search alongside ScyllaDB Vector Search in order to bring both lexical and semantic retrieval closer to operational data.

Built for Performance

ScyllaDB Full-Text Search is designed for latency-sensitive, high-throughput workloads.

Full-text indexes run on dedicated indexing nodes and are maintained in memory for fast retrieval. This architecture lets teams add search capacity without changing the database tier, while continuing to support heavy write workloads and rapidly changing datasets.

ScyllaDB CEO Dor Laor said, “Full Text Search extended our ScyllaDB X Cloud offering, bringing ScyllaDB superior performance and elasticity to the Text Search domain. Both Text and Vector Search are important parts of our realtime AI offering.”

P99 CONF, a free + virtual conference on “all things performance,” will feature a talk on Benchmarking ScyllaDB Native Full-Text Search vs. OpenSearch. The conference also includes 60+ other tech talks by engineers from Canva, Palo Alto Networks, Uber, Netflix, Lovable, Yahoo, DoorDash, Disney/Hulu, Snap, American Express, and more. Free conference passes are available at p99conf.io.

Full-Text + Vector Search

Full-Text Search complements ScyllaDB Vector Search.

Vector Search helps applications find content with similar meaning. Full-Text Search helps find documents containing specific terms and phrases.

Applications can use both on the same underlying ScyllaDB data and combine the results for hybrid retrieval, making this especially useful for RAG, AI agents, support systems, product catalogs, and technical search.

Available in ScyllaDB 2026.3

ScyllaDB 2026.3 Full-Text Search includes:

• BM25 relevance ranking
• Boolean and phrase search
• Language-specific analyzers and stemming
• Multiple full-text indexes per table
• Full-text and vector indexes on the same table
• CQL-native querying
• Dedicated, independently scalable indexing nodes
  
  

About ScyllaDB

ScyllaDB is a specialty database for workloads that require predictable performance at scale. It’s adopted by organizations that require ultra-low latency, even at millions of features or operations per second, billions of embeddings, or petabytes of storage. ScyllaDB’s shard-per-core architecture taps the full power of modern infrastructure, translating to fewer nodes, less admin, and lower costs. Over 400 companies such as Disney+, Discord, Tripadvisor, Expedia, Zillow, Starbucks, and Comcast use ScyllaDB for their toughest database challenges. For more information: https://www.scylladb.com/

Media Contact:
Wayne Ariola
wayne.ariola@scylladb.com