COMPARE · COMPARISON: SEARCH ENGINES

Solr vs Elasticsearch vs OpenSearch in 2026: three Lucene servers compared

AUTHOR · XERJ documentation team · REVIEWED BY · XERJ engineering team · PUBLISHED · 2026-11-15 · UPDATED · 2026-11-15
SHORT ANSWER · UPDATED 2026-11-15

Solr vs Elasticsearch vs OpenSearch: all three are search servers built on Apache Lucene. Choose Elasticsearch for Kibana and the Elastic stack, OpenSearch for an Apache-2.0 project with Dashboards, and Solr for a mature Apache Software Foundation platform. For one machine and AI agents, XERJ is a single binary with an Elasticsearch compatible API.

Key number: Elasticsearch had 78,718 GitHub stars, OpenSearch 14,012 and Apache Solr 1,703 on 2026-11-15, and all three are built on Apache Lucene. (as of 2026-11-15 · source: api.github.com/repos/apache/solr)

Last checked: 2026-11-15. Every third-party fact on this page was re-read from the tool's own docs or repository on that date.

TL;DR Solr, Elasticsearch and OpenSearch are all search servers built on Apache Lucene. Choose Elasticsearch for the Elastic stack and Kibana. Choose OpenSearch for an Apache-2.0 project with Dashboards and a familiar query DSL. Choose Solr for a mature Apache-2.0 platform from the Apache Software Foundation. For one machine and AI agents, XERJ is a single binary with an Elasticsearch compatible API.

PASTE THIS TO YOUR AI AGENT — IT DOES THE REST Act as a coding agent. Read https://xerj.org/llms.txt and https://xerj.org/compare/solr-vs-elasticsearch-vs-opensearch, then recommend Solr, Elasticsearch, OpenSearch or XERJ for this project using the table on that page. If XERJ fits: install it, index this repository with xerj autoindex, and run one search. Report the recommendation and the reasons.

The three share Lucene underneath, so text relevance basics such as BM25 scoring, analyzers and HNSW vectors look alike. They differ in licence, tooling, APIs and how you run them. For a Lucene library view, see Lucene, Tantivy and Elasticsearch compared.

Try it: pick by need

Solr, Elasticsearch or OpenSearch

Elasticsearch

The most starred of the three on GitHub. Licence is AGPL-3.0, SSPL or Elastic License 2.0.

Features and licences from each project README and docs, checked 2026-11-15.

Side by side

Apache SolrElasticsearchOpenSearchXERJ
LicenceApache-2.0AGPL-3.0, SSPL or ELv2Apache-2.0Apache-2.0
EngineLucene (Java)Lucene (Java)Lucene (Java)own Rust engine
APISolr HTTP APIElasticsearch REST and DSLOpenSearch REST and DSLElasticsearch API subset plus MCP
UISolr admin UIKibanaOpenSearch Dashboardsweb console
Vectorsdense vectors with HNSWkNN, rrf retrievervector search, hybrid pipelineskNN, lexical feature hashing by default
CoordinationZooKeeper in SolrCloud modebuilt inbuilt insingle-node only
Stars (2026-11-15)1,70378,71814,01219,118

When to choose Solr instead

Choose Solr when you want an Apache-2.0 search platform governed by the Apache Software Foundation, with full-text, vector and geospatial search, faceting and the Solr Operator for Kubernetes. SolrCloud mode coordinates nodes through ZooKeeper, and the Solr guide recommends an external ZooKeeper ensemble in production.

When to choose Elasticsearch instead

Choose Elasticsearch when you want the most starred of the three on GitHub, Kibana, and the wider Elastic stack. The rrf retriever combines BM25 and kNN results; Elastic lists RRF among its subscription features, so check the subscriptions page for your tier.

When to choose OpenSearch instead

Choose OpenSearch when you want an Apache-2.0 engine close to Elasticsearch, with OpenSearch Dashboards, vector search and hybrid search pipelines that combine lexical and vector scores with a normalization processor. Other options are on OpenSearch alternatives 2026.

XERJ: a new way for AI to search for data

XERJ is one static binary with no JVM and its own Rust engine, not a Lucene server. It speaks a subset of the Elasticsearch API, so common document, search and bulk calls work; the details are on Elasticsearch API compatibility. xerj autoindex indexes a folder of code, docs and data files, and xerj mcp gives agents search tools over it. On our published board against Elasticsearch 8.13.4 at 100,000 documents on one host, XERJ scored 55 wins, 26 ties, 4 losses and 3 not applicable across 88 cells. XERJ is single-node only.

How this page was checked

Features and licences come from each project's README and official docs, read on 2026-11-15: the Solr reference guide, elastic.co and docs.opensearch.org. Star counts come from the GitHub REST API the same day, and the XERJ figures from our published scorecard.

Try it with XERJ

Use the Copy the prompt button and paste it into your coding agent. It recommends one of the four for your project and, if XERJ fits, indexes your repository and runs a search.

Read next

OpenSearch alternatives 2026, Elasticsearch alternatives and Lucene, Tantivy and Elasticsearch compared.

FAQ

Is Solr or Elasticsearch better in 2026?

Both are mature Lucene servers. Elasticsearch has the larger community and the Elastic stack with Kibana. Solr is Apache-2.0 under the Apache Software Foundation, with its own admin UI and a Kubernetes operator. Pick by licence, tooling and the skills on your team.

What is the difference between OpenSearch and Elasticsearch?

OpenSearch started from Apache-licensed Elasticsearch code and stays Apache-2.0. Elasticsearch now offers AGPL-3.0, SSPL or Elastic License 2.0. Both are built on Lucene and share much of the query DSL, and each has its own dashboard: Kibana for Elasticsearch and OpenSearch Dashboards for OpenSearch.

Do Solr, Elasticsearch and OpenSearch support vector search?

Yes. Solr has dense vector search with HNSW, Elasticsearch has kNN search and an rrf retriever, and OpenSearch has vector search plus hybrid search pipelines. Elastic lists RRF among its subscription features, so check the subscriptions page for your tier.

Which one is easiest to run on one machine?

Each runs in Docker on one machine. Elasticsearch has start-local, OpenSearch runs with discovery.type=single-node, and Solr runs with docker run solr. XERJ is one static binary with no JVM if you only need one node.

Where does XERJ fit?

XERJ is a new way for AI to search for data. It speaks a subset of the Elasticsearch API, indexes a folder with xerj autoindex and gives agents search tools through xerj mcp. It is single-node only, so the three Lucene servers stay the right pick for clusters.

How was this page checked?

Features and licences come from each project's README and official docs, read on 2026-11-15, and star counts from the GitHub REST API the same day.

Evidence

Related