// the find
typesense/showcase-recipe-search
Instantly search 2M cooking recipes using Typesense Search (an open source alternative to Algolia / ElasticSearch) ⚡ 🥘 🔍
A demo app showing Typesense search across 2.2M recipes, built with the Typesense InstantSearch adapter. It's a showcase project, not a production template — the point is to demonstrate what Typesense can do at scale, not to give you reusable recipe-search code.
The indexing numbers are real and stated plainly: 2.2GB dataset indexed in 8 minutes on a 3-node cluster, 2.7GB RAM footprint. That's useful benchmark data if you're evaluating Typesense for a similar scale. The InstantSearch adapter integration is clean and minimal — good reference for how little glue code you actually need. The geo-distributed cluster setup (Oregon, Frankfurt, Mumbai) is a concrete example of Typesense Cloud's multi-region story, not just a claim.
The indexer is Ruby, the frontend is JS — two runtimes for a demo is unnecessary friction and Ruby 2.7.2 is EOL. If you're trying to replicate this locally you're setting up a second toolchain just to parse CSVs. There's no schema definition visible in the repo structure, so you can't see how the collection is configured without running the indexer first. The entire thing points at Typesense Cloud, so the 'self-host this' angle requires more work than the README implies — the local Docker path exists but isn't the main story. No tests anywhere.