Elasticsearch Learning To Rank

This plugin brings learning to rank into Elasticsearch, so result ordering can be driven by a trained model instead of hand tuned query boosts. Features are stored inside Elasticsearch as query templates, their scores are logged alongside relevance judgments to build a training set, a model is trained offline from that set, and the stored model then reranks the top results of a query. It is aimed at search engineers and relevance teams who already run Elasticsearch and want to move from manual tuning to a measurable, model driven approach. Linear, XGBoost and RankLib models are supported, so training can be done with familiar ranking libraries outside the cluster. Feature logging is the part that usually matters most in practice, because it produces training data from the same queries production is already serving. The plugin was created by OpenSource Connections, and a separate demo project walks through building a feature set, logging scores, training a model and deploying it. The plugin is free and open source under the Apache 2.0 license. There is no paid edition or hosted service. The maintainers note that paid public training courses on relevance help fund the free development work. One practical constraint is version binding. Each release is built against an exact Elasticsearch version and installed from a prebuilt archive matched to it, and the maintainers state that they aim to support point one releases and ask users who need another build to contribute it. The repository remains active, with commits through 2026.