Typesense

Typesense is an open-source search engine from Typesense, Inc. built for instant, typo-tolerant search, with a RESTful API and client libraries for many languages. It can be self-hosted via Docker, native binaries, DEB/RPM packages or Homebrew, or run on the managed Typesense Cloud. The latest documented server version is 30.2.

What It Is

Typesense is a search engine that developers use to add search-as-you-type, autocomplete, faceted navigation and similar experiences to apps and sites. You create a collection with a schema, import documents, and query them through the HTTP API or a client library. The company positions it as an open-source alternative to Algolia and an easier-to-use alternative to Elasticsearch.

Search Capabilities

The documentation lists typo tolerance, tunable query-time ranking, synonyms, dynamic sorting, merchandising (pinning records), grouping and distinct, filtering and faceting, geo search, and federated multi-search. It also supports vector and semantic search, built-in RAG, and recommendations. Multi-tenant scoped API keys allow data for multiple users to live in one collection with access rules.

Deployment and Operations

Typesense runs from Docker images, Linux binaries, DEB or RPM packages, or Homebrew. Replication provides high availability. Typesense Cloud provisions clusters, and the docs describe an MCP server through which an AI agent can provision a cluster, index data and run searches.

Ecosystem and Integrations

Official and community client libraries cover JavaScript, Go, PHP, Ruby, Python, Dart, .NET, Java, Swift and Clojure. Integrations include Laravel Scout, WordPress, Gatsby, Firebase, LangChain, Craft CMS, WooCommerce, and UI adapters for InstantSearch.js, DocSearch and Docusaurus. The typesense-js client works server-side and in the browser, and the docs advise using a search-only API key in browsers.