Meilisearch: Search That Works Out of the Box

A review of Meilisearch — an open-source search engine with instant search, typo tolerance, and a simple REST API for any project.

Meilisearch: A Search Engine That Doesn’t Annoy You

Meilisearch is an open-source search engine written in Rust. Its main selling point is “it just works”: setup takes minutes, and search results are relevant without fine-tuning ranking.

Why Meilisearch

Most search solutions (Elasticsearch, Solr) require serious configuration and administration. Meilisearch took a different path:

  • Instant search (search-as-you-type) — results appear on the first keystroke
  • Typo tolerance — “kat” finds “cat” without additional configuration
  • Filters and facets — all through the REST API
  • Relevance sorting — customizable ranking rules
  • Excellent documentation — interactive examples, SDKs for all popular languages

Technical Details

Meilisearch runs as a separate service. Data indexing is done via the REST API: you send JSON documents, and you get a search index. Under the hood — a custom engine in Rust, delivering speed and low memory consumption.

It supports: synonyms, stop words, hit highlighting, geosearch. Russian language is supported out of the box.

Limitations

  • Maximum index size — around 100 million documents (for very large volumes — Elasticsearch)
  • No distributed search in the open-source version
  • The open-source license changed from MIT to proprietary in 2023 (but older versions remain MIT)

Verdict

For most projects — websites, e-commerce stores, documentation sites — Meilisearch delivers the ideal balance of simplicity and search quality. I recommend starting with it instead of Elasticsearch.

Official Meilisearch website