OpenSearch Project Blog

When discussing performance optimizations in OpenSearch, we usually focus on the processing of actual data: we look at algorithms and data structures to improve the amount of time a query...

Author Daniel Wrigley Daniel is a Search Consultant at OpenSource Connections. He has worked in search since graduating in computational linguistics studies at Ludwig-Maximilians-University Munich in 2012 where he developed...

Authors Shenoy Pratik Gurudatt Shenoy Pratik is a software engineer at AWS working on observability for the OpenSearch Project. He maintains multiple plugins, including OpenSearch Observability, Reporting, and Query Workbench....

In a recent episode of the OpenObservability Talks podcast, I sat down with Carl Meadows, Chair of the Governing Board of the OpenSearch Software Foundation, and Pallavi Priyadarshini, Technical Steering...

Why Retrieval-Augmented Generation (RAG) matters Large language models (LLMs) have transformed AI with their ability to generate human-like text, but they have notable limitations. Because an LLM’s knowledge comes from...

Learn how OpenSearch improves the performance of date histogram aggregations by using the BKD index tree, with support for sub-aggregations and multi-range traversal.

Semantic search improves result relevance by using a machine learning (ML) model to generate dense or sparse vector embeddings from unstructured text. Traditionally, enabling semantic search has required several manual...

This blog provides an overview of the OpenSearch User Group program and some observations I shared as part of a keynote presentation at this year’s OpenSearchCon Europe. The OpenSearch Project has grown...

In today’s data-rich world, information isn’t just text anymore. It’s images, videos, audio, and more. To truly understand and retrieve relevant information, search engines need to go beyond keywords and...

At KRUU, Europe’s leading photo booth rental provider, smooth logistics are critical to customer satisfaction. The company’s signature product, a rentable, shippable photo booth for weddings and special events, is...

Traditional lexical search, based on term frequency models like blend term frequency models (BM25), is widely used and effective for many search applications. However, lexical search techniques require significant investment...

The Approximation Framework is a performance optimization in OpenSearch that introduces early termination during BKD tree traversal. It enhances the efficiency of numeric range, sort, and match_all queries by collecting only the required number of documents. This framework delivers exact results with significantly reduced query latency.

Explore advanced semantic field features in OpenSearch, including text chunking, support for externally hosted and custom models, cross-cluster usage, and updating model references.

Explore OpenSearch 3.0's semantic highlighting, an AI-powered feature that intelligently identifies relevant passages based on meaning rather than exact keyword matches. Learn how to enhance your search applications with this powerful capability.

Discover OpenSearch 3.1's system ingest pipeline feature, enabling plugin developers to automate document processing during ingestion. Learn how it simplifies configuration, improves user experience, and enhances data transformation capabilities.

Learn how to use ML inference processors to implement semantic search and summarization in OpenSearch.

OpenSearch 3.1 delivers up to 3.8x faster hybrid query throughput by optimizing score collection, batching, and segment-level execution—resulting in lower latency and more predictable performance under load.

Discover how to implement neural sparse models in Sentence Transformers with OpenSearch. Learn to encode text into sparse vectors with just a few lines of code for efficient search.

OpenSearch Data Prepper 2.12 introduces new features to enhance your data ingestion and processing pipeline. Discover the unified OTLP source for streamlined telemetry data ingestion, the Amazon SQS sink for efficient message queuing, and integration with AWS X-Ray for observability.