Skip to content

Key Features

Nixiesearch is built to make search simple, scalable, and safe - especially compared to the complexity of running Elasticsearch or OpenSearch in production. Here are a few of the standout features that make Nixiesearch different:

đź”’ Schemaless is not OK

Unlike databases, where schemaless design often leads to chaos, many search engines (we're watching at you, Elasticsearch) still let you get away with it, until it breaks everything.

Why it matters:

  • Mapping changes can wreck your index: In Elasticsearch, altering a mapping in an incompatible way while running indexing can silently break your queries, when half of documents are processed in a new way, and half of them stays in an old format.

  • Nixiesearch enforces document schema on ingestion, and validates correctness of schema migrations (so you cannot flip field type from int to text without explicit reindexing). You always know exactly what’s indexed and how. No surprises, no “uh-oh” moments.

Wildcard support is still there when you really need flexibility, but only in a controlled and explicit way.

📦 Immutable S3-Based Indexes

Storing indexes on S3 isn't just a cost-saver - it changes the game operationally.

What’s different:

  • Truly immutable index files: No risk of corruption or half-written segments. Snapshots and backups are just… copies on S3.
  • No cluster state: All config lives in plain text. Upgrades, schema changes, and rollbacks are simple file updates. You can GitOps your whole search setup in a simple ArgoCD manifest.
  • Zero-downtime everything: Want to change schema? Just deploy a new config and do a rolling or blue-green restart of the Kubernetes Deployment. No need to babysit an existing cluster.

🧲 Pull-Based Indexing

Indexing should never interfere with your search workload.

What’s different:

  • Nixiesearch uses a pull-based architecture: indexing runs in a separate tier, so your search nodes never get overloaded.
  • Supports both offline batch (for reindexing) and real-time streaming (Kafka, S3, etc.), so you can scale indexing independently.
  • You get natural backpressure and safe ingestion by design, without needing to tune queues or buffers.

🧠 Power of Apache Lucene

At its core, Nixiesearch runs on Apache Lucene — the same battle-tested engine behind Elasticsearch, but without the operational baggage.

What you get:

And yes, it's fully open-source—no feature gating or “you need Enterprise” surprises.

Whether you're migrating from Elasticsearch or starting fresh, Nixiesearch offers a radically simpler way to build production-grade search. It's the search engine that feels like infrastructure should in 2025: immutable, stateless, scalable, and smart.