orama

Why we included this project

Orama runs a full search engine inside your application process instead of forcing you to spin up a separate service. It's a TypeScript library that handles full-text, vector, and hybrid search in the browser, on a Node server, or on edge runtimes, with a claimed footprint under 2kb. That makes it a natural fit for client-side search over local datasets and offline-first apps, where a dedicated database would add latency and ongoing operational overhead. The package ships the practical features developers reach for quickly: BM25 ranking, typo tolerance, filtering, facets, field boosting, geosearch, and stemming in 30 languages. It also includes a RAG pipeline for grounding LLM answers in your indexed content, so teams can go from indexing to retrieval-augmented chat without stitching together several libraries.