Typesense Search Primer

Why Typesense

Typesense is an open-source, in-memory search engine built for speed. It delivers sub-50ms results with typo correction, faceting, and geosearch out of the box. A single binary replaces the sprawling Elasticsearch stack — ideal for SaaS docs, e-commerce SKU search, or any user-facing lookup.

Core Concepts

• Collections — analogous to SQL tables. Define a schema with fields like string, int32, or float.
• Documents — JSON objects you index. Each doc gets a unique id.
• Search parameters — q for query, query_by for fields, filter_by for faceted constraints.

Quick Start

Run Typesense locally via Docker or use Typesense Cloud for a managed cluster. Install the official typesense-js client, create a collection schema, index your documents, and call client.collections('docs').documents().search(). For Next.js App Router, wrap calls in a Route Handler to keep your API key server-side.

Nimbus Integration

Meridian's docs search uses Typesense to index every recipe, guide, and API reference. Typo-tolerant matching means “typsense” still finds this page. Faceted filters let you narrow by category or difficulty tier. The search bar in the top nav hits a serverless endpoint backed by Typesense Cloud — cold starts are under 80ms.