Powering the generative AI era

The Elasticsearch Relevance Engine™ (ESRE) is designed to power artificial intelligence-based search applications. Use ESRE to apply semantic search with superior relevance out of the box (without domain adaptation), integrate with external large language models (LLMs), implement hybrid search, and use third-party or your own transformer models.

A relevance engine tailor-made for developers who build AI powered search applications.

Read launch blog

Learn everything about Elastic's new retrieval model that delivers high relevance, even on specialized domains in a zero-shot setting.

Read tech blog

In our investor keynote, discover why Elasticsearch is a critical part of building Generative AI experiences.

Watch video


Subscribe for AI updates

Stay informed on all things related to AI. Sign up to receive exclusive news, Elastic® product updates, AI trends, hands-on demos, and more.


Straight from the source

LLMs, NLP, ChatGPT, RRF, BM25, ESRE, vectors, transformers, embeddings — there’s a lot to keep up with in the world of AI. Hear from Elastic product and engineering leadership and development teams about this fast-moving AI landscape so you can build better AI-powered search applications.

Sign up


AI for all developers

Elevate search with AI

Create innovative search applications, unhindered. With the Elasticsearch Relevance Engine, you get speed, scale, and relevance with flexibility and control to deploy machine learning and Generative AI search apps however you see fit.

  • Supercharge search relevance

    Use your own domain-specific data store in context windows to improve the relevance of LLMs' human-like output. Get best-in-class semantic search using Elastic Learned Sparse Encoder that can be deployed in one click. No training or model upkeep necessary — it's ready to use.

  • Generate embeddings & search vectors at scale

    Convert unstructured data into vector embeddings, efficiently search them using approximate nearest neighbor search, and optimize storage and search latency with quantized byte vectors. Elasticsearch brings all the capabilities you need to build semantic search or any vector search application.

  • Bring your own transformer models

    Use the Eland Python library to bring your own trained machine learning model, or use a third-party model from a public repository. Pick from a wide variety of supported architectures that fit your needs.

Elasticsearch Relevance Engine

Generative AI, with the Elasticsearch advantage

Generate embeddings. Store and search vectors. Get semantic search with Elastic's own Learned Sparse Encoder machine learning model. Ingest all data types. Integrate with rapidly evolving large language models.

  • Vector database

    Get a full vector search experience – don’t just store and search embeddings, create them! Capture the meaning and context of your unstructured data, including text and images, with embeddings for dense retrieval. Use vector search as a starting point to integrate with Generative AI.

  • Elastic Learned Sparse Encoder

    Our new model delivers highly relevant semantic search out of the box, without domain adaptation. It’s available with one click while configuring your search application. Elastic Learned Sparse Encoder expands queries with related keywords and relevance scores, so they’re easily interpretable and ready for use right away.

  • Large language models

    Give LLMs business-specific information using your private data (not just publicly trained data). Use Elasticsearch for high relevance context windows that draw on your proprietary data to improve LLM output and relevance. Access Generative AI with APIs and plugins integrated with the LLM of your choice.

  • RRF hybrid ranking

    RRF (Reciprocal Rank Fusion) is a method for combining document rankings from multiple retrieval systems. In the near future, RRF will support blending results from sparse vector models like BM25 and Elastic's retrieval model yielding the best in class zero shot ranking method. Hybrid ranking with RRF lets you tune search results from multiple retrievers with less effort.

  • Transformer models

    Bring your own proprietary transformer model into Elastic. Or upload pretrained models from third-party repositories like the HuggingFace model hub — with support for a variety of supported architectures such as BERT, BART, ELECTRA, and more.

  • Data integrations & ingestion libraries

    Familiar tools such as Elastic Agent or Logstash to index your data. An ever-expanding list of integrations (such as Confluence, S3, or Google Drive). Native database connectors (such as MySQL, MongoDB). A web crawler for online sources. For custom app data, Kibana APIs or build-your-own connector with familiar frameworks.

Frequently asked questions

What is Elasticsearch Relevance Engine?

Elasticsearch Relevance Engine is a set of tools that help developers build AI search applications and includes:

  • Industry leading advanced relevance ranking features, including traditional keyword search with BM25, a foundation of relevant, hybrid search for all domains.
  • Full vector database capabilities – including the ability to create embeddings, in addition to storage and retrieval of vectors.
  • Elastic Learned Sparse Encoder – our new machine learning model for semantic search across a range of domains.
  • Hybrid ranking (RRF) for pairing vector and textual search capabilities for optimal search relevance across a variety of domains.
  • Support to integrate 3rd-party transformer models such as OpenAI GPT-3 and 4 via APIs.
  • A full suite of data ingestion tools such as database connectors, 3rd-party data integrations, web crawler, and APIs to create custom connectors.
  • Developer tools to build search applications across all types of data: text, images, time-series, geo, multimedia, and more.