Skip to Content
Relevan Beta is here 🎉
ResourcesCompareRelevan vs OpenSearch

Relevan vs OpenSearch

OpenSearch is an open-source search engine. You can run it yourself or use a managed service, such as Amazon OpenSearch Service. Relevan runs on OpenSearch, and it adds the relevance work for each index and an API for agents.

The facts about OpenSearch come from its public documentation and the AWS documentation, as of October 2026. The sources are at the end of this page.

Summary

OpenSearchRelevan
ApproachSearch infrastructureRelevance tuning for each index
LicenseOpen source, Apache 2.0Hosted service on OpenSearch
Who runs the clusterYour team, or a managed serviceRelevan
Who tunes relevanceYour teamRelevan
Click and feedback dataA schema and plugins. Your team builds and runs the pipeline.An API, with logging and storage
PricingYour servers, or the price of the managed serviceMonthly plans, from a free plan

What OpenSearch does

  • OpenSearch gives full-text search with BM25, vector search, hybrid search, filters, and aggregations.
  • The Learning to Rank plugin rescores results with XGBoost or RankLib models. On Amazon OpenSearch Service, you make the judgment list, prepare the training data, and train the model outside the service.
  • User Behavior Insights (UBI) is a schema for searches and the actions that follow them. An optional plugin and a JavaScript library help you collect the data.
  • The Search Relevance Workbench helps you compare queries, judge results, and run A/B tests.

What Relevan does

Relevan is a managed search engine for your own records, such as products, profiles, and documents. You send JSON records and a mapping. Relevan runs the OpenSearch index, and agents search it through an HTTP API.

  • Each search response tells the agent why each result matched, and suggests what to try next.
  • Relevan tunes retrieval and ranking for each index. Compare Relevan with other search tools explains what Relevan tunes, and when.
  • The feedback endpoint records clicks, LLM judgments, and conversions for each search.
  • Relevan prices by plan, from a free plan. See pricing.

The main difference

OpenSearch gives you each part of a search system, and your team puts the parts together. Relevan uses OpenSearch as its engine. Relevan sets the field weights, the BM25 parameters, and hybrid retrieval for each index. It can also add a reranking model for each index.

Click and feedback data

To train a ranking model on OpenSearch, your team builds and runs the path for click and feedback data. These are the parts:

  1. A log of each search, with a query ID, the results, and the position of each result.
  2. An endpoint or a client library that collects clicks, judgments, and conversions.
  3. A stream or a queue that carries the events, with buffering, retries, and alarms.
  4. Storage, partitioned by customer, index, and date.
  5. A retention policy, and the removal of personal data.
  6. A join of each event to the search that it is about.
  7. A weight for each event type. A click is weaker evidence than a conversion.
  8. A job that turns events into judgment lists, with a correction for position bias.
  9. Training, evaluation, and deployment of the model, and a new run when the data or the mapping changes.

Relevan supplies parts 1 to 4 and part 6 through its API, and the retention window of part 5. Each search response has a query ID, and the feedback endpoint takes events that refer to it. Relevan logs the searches and the events, and stores them by organization, index, date, and event type. Relevan keeps these events for 13 months by default. Click and feedback data gives the details, and the part that is not automatic today.

When to choose OpenSearch

  • You have a team that can run a cluster and tune relevance.
  • You need features that Relevan does not expose, such as aggregations of your own design, custom analyzers, or log analytics.
  • You want to control the version, the plugins, and the hardware.
  • You must run the engine in your own account or data center.

When to choose Relevan

  • You want OpenSearch search with no cluster to size, shard, or upgrade.
  • You do not have a team to build the click and feedback pipeline.
  • You want each index tuned to its own documents and queries.
  • Your agents need an explanation of each match and a suggestion for the next search.

Sources

Read on October 29, 2026:

Last updated on
Relevan vs MixedbreadRelevan vs turbopuffer