• Learning To Rank (LTR): Native plugin for ML-based relevance optimization using XGBoost and RankLib models. Takes click and conversion signals into account when ranking.
• Security Plugin: Includes advanced security (RBAC, encryption, audit logs) for free, which are paid features in the Elastic Stack.
• SQL / PPL Querying: Allows querying data using familiar SQL or Piped Processing Language syntax.
• Neural Search: Built-in support for ML models to run semantic search pipelines directly within the cluster.
• Trace Analytics & Observability: Strong observability features for distributed application tracing, dashboards, and time-series data.

OpenSearch is the open-source challenger. Positioned as a fully open alternative to Elasticsearch, it has spread rapidly and, with more than 1 billion downloads reported in 2025, is one of the fastest-growing search technologies around. Elastic still tends to lead on brand-new features and on the most polished hybrid developer experience, but OpenSearch is the obvious choice for anyone heavily invested in AWS.
It is no longer purely an AWS story. Other providers offer OpenSearch as a managed service across various cloud platforms, and it can be self-hosted on-premise without dependence on a single vendor. Governance now sits with the vendor-neutral OpenSearch Software Foundation rather than one company. This shows that OpenSearch has moved well beyond being a mere Elasticsearch derivative and is evolving independently.