AI Search
Real-time search and matching for production applications
AI-powered hybrid search for real-time applications
Hybrid semantic and keyword search, fully managed and governed by Unity Catalog
Power real-time production applications with Databricks AI Search, a fully managed search engine built on the Databricks Platform. Combine semantic, keyword (BM25) and hybrid search in one API, with built-in reranking and quality evaluation. Automated sync pipelines keep your indexes as fresh as your source data changes, with no pipelines to maintain.
Unity Catalog applies your existing access controls and lineage automatically. Serverless from day one, it scales to billions of records and thousands of queries per second so your apps stay fast as they grow.
Hybrid search in one API
Combine semantic vector search with keyword search (BM25) in a single call, then rerank for production-grade relevance. Power product search, content recommendations, e-commerce experiences and agents on one engine, with no separate systems to operate.
Real-time data sync, with no pipelines to manage
Our automatic sync pipelines update your indexes as source tables change. Databricks handles embedding generation, retries, throughput tuning, batch sizing and autoscaling, so your search stays in step with your data without any data engineering work.
Governance built in
AI Search inherits the access controls, lineage and audit logs you've already defined in Unity Catalog. Fine-grained policies apply to indexes as a Unity Catalog object.
Built for production scale
Fully managed and serverless, AI Search scales to billions of records and thousands of queries per second. No clusters to size, no shards to balance, no replicas to plan — just consistent low-latency search as your traffic grows.
