Typesense Search Engine

Typesense is an open-source, typo-tolerant search engine created by Jason Bosco in 2019, written in C-plus-plus, providing fast, relevant search with features including typo-tolerance, synonyms, curating, geo-search, federated search, vector search, and multi-tenant search, designed as an alternative to Algolia with an emphasis on being lightweight, easy to operate, and cost-effective, with the ability to tune relevance at query time without reindexing, adopted by over twenty thousand GitHub stars and companies including Crunchbase and Day One Journal for their search infrastructure. The typo-tolerance system provides configurable spelling correction with parameters controlling the maximum number of typos allowed and when to fall back to exact matches, with the engine supporting prefix search for instant results as the user types, and word-level correction that handles misspellings in any word of the query. The vector search capability enables semantic similarity search through embedding vectors, with the engine storing dense vector representations alongside text fields, enabling hybrid search that combines traditional full-text relevance with vector similarity for more nuanced result ranking, useful for recommendation systems and natural language search. The geo-search enables location-based filtering using latitude and longitude coordinates, with configurable radius filtering and distance-based sorting, ideal for store locators and local search. The multi-tenant search enables per-tenant result customization through API key-based filtering. The federated search across multiple collections. The curation for pinning specific results. The synonym management. The self-hosted and cloud options. Designed by Jason Bosco. Designed for developers.