Meilisearch

Meilisearch is a search engine written in Rust that has accumulated over 47,000 GitHub stars by making the "add search to your app" experience as close to painless as the category allows. The core is MIT-licensed, which means you can run it commercially without license negotiation. Deployment is a single binary or Docker image that exposes a clean JSON REST API — indexes are created by posting your documents, and search is immediately available with typo tolerance, prefix matching, and faceting working out of the box without configuration.

The relevance defaults are deliberately opinionated and generous. A query like "restrant" against a food database of restaurants will return the right results because Meilisearch applies configurable edit-distance tolerance per field. Custom ranking rules, stop words, synonyms, and geo-search are all built in and set through the same REST API or any of the official SDK clients (JavaScript, Python, PHP, Ruby, Go, Rust, Swift, Dart, Java). The developer experience has driven adoption among teams that previously found Elasticsearch operationally heavy for straightforward search use cases.

The architecture trade-off is RAM. Meilisearch keeps its indexes in memory-mapped files for performance, which means the machine running it needs enough RAM to hold the working index. Large datasets — multi-million document indexes — require planning around memory capacity and potentially horizontal clustering via its experimental multi-node support. The Meilisearch Cloud managed service is available for teams that want the developer experience without the infrastructure work, though that reintroduces usage-based pricing.