From App Search to Elasticsearch — Tap into the future of search

Recent advancements in generative AI are transforming user behavior, inspiring developers to create search experiences that are more dynamic, intuitive, and engaging. At Elastic, we’re focused on equipping developers with powerful machine learning (ML) tools in Elasticsearch to push the boundaries of modern search experiences.

As part of our commitment to innovation, we are making an important transition.  

• We’ve packed Elasticsearch with integrated search and machine learning tools, making semantic search — powered by Elastic Learned Sparse EncodeR (ELSER) — as simple as a single field type definition.

• We're simplifying the architectural choices developers need to make by discontinuing App Search in 9.0.

• We’ve put migration on your timeline: App Search will remain with its current feature set in the 8.x series, and we will continue to provide security upgrades and fixes.

For users new to the Elasticsearch capabilities, the same ease-of-use and out-of-the box search functionality that App Search users have traditionally enjoyed are all now integrated into the Elasticsearch experience. Now, users can have it all — from an accessible start that gets you searching within minutes to an infinitely customizable set of search tools that can be fine-tuned to the specifics for your use case.

Here's what you can expect:

• Semantic search made simple: 

  • The new semantic_text field and semantic query allow for ML-powered semantic search with just a single field.

  • An out-of-the-box sparse vector model (ELSER) for semantic search or the choice to bring your own

• Enhanced relevance tools: 

  • Improve your relevance with mid- and late-stage reranking models with semantic reranking and a native implementation of Learning to Rank.

  • An out-of-the-box cross encoder reranking model (Elastic rerank) for semantic reranking or the the choice to bring your own 

• Powerful vector capabilities: Access vector database and vector search tools, and easily combine vector and token search with hybrid techniques.

• State-of-the-art vector data compression techniques: Check out BBQ! 

• Large language model (LLM)-powered chat experience: Get your retrieval augmented generation (RAG) workflows started quickly with an out-of-the-box chat experience powered by LLMs with AI Playground!

• Streamlined architecture: Eliminate the need for Enterprise search nodes while enabling efficient scaling and delivering performance boosts through index tuning and optimized Elasticsearch queries.

All of the above — with great UI experiences to manage your relevance — measure the efficacy of your search and expand to future goals that meet our organization’s search needs.

The release and timing of any features or functionality described in this post remain at Elastic's sole discretion. Any features or functionality not currently available may not be delivered on time or at all.