Typesense

Lightning-fast, typo-tolerant, open-source search engine optimized for instant, relevant results. In-memory architecture delivers search-as-you-type with minimal configuration, combining full-text search with vector and semantic retrieval in a REST API designed for quick deployment.

Trusted by teams at Codecademy, Logitech, BBC Maestro, Kick, Lonely Planet, and ElevenLabs. Typesense positions itself as an open-source alternative to Algolia and a simpler-to-operate option than Elasticsearch: no complex cluster tuning, typo tolerance by default, and scoped API keys for multi-tenant indices. Official client libraries cover JavaScript, Python, Go, PHP, Ruby, Swift, Java, Dart, and .NET.

Key capabilities:

• Typo-tolerant full-text search with tunable ranking, synonyms, and merchandising
• Vector and semantic search for recommendations, RAG, and meaning-based retrieval
• Faceted navigation, geo search, and federated search across multiple indices
• Multi-tenant API keys with per-collection access control
• Geo-distributed deployment for low-latency reads close to users
• Integrations with Laravel Scout, Gatsby, WordPress, Firebase, Craft CMS, and LangChain

Common setups include e-commerce product search with faceted filters, autocomplete and search-as-you-type for content sites, geo-search for local listings and store finders, and vector-backed RAG pipelines for chat applications. Developers often sync from a primary database into Typesense and use it as a secondary search layer rather than a primary data store.