Lyra: Disrupting full-text search industry with JavaScript
summary
A JavaScript full‑text search engine claims Lucene‑grade speed while running anywhere JS runs and reducing deployment and cost pain. It indexes 20,000 products in ~3 seconds in the browser, serves queries from a CDN at 72–702 microseconds, and supports facets, filters, typo tolerance, plugins, hooks and custom indexes. It runs on Node Deno Cloudflare Workers and browsers, exposes opt‑in data structures, and offers precomputed deltas and LLM integration to tailor results.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 0:48Elasticsearch inspired Lyra
- 3:00Search engine challenges
- 4:03Goals for new engine
- 5:05JavaScript over Rust/Go
- 6:29Array intersection optimizations
- 8:16JavaScript performance myths
- 10:26V8 optimizations pitfalls
- 11:29JavaScript in microservices
- 12:12Prototype to open source
- 13:38Founding Orama team
- 14:19Orama setup simplicity
- 15:45Inserting querying data
- 16:47Orama unique features
- 18:13Live demo indexing
- 20:24Cost efficiency benefits
- 21:49LLM integration with Orama
- 22:31Call to action hiring
- 23:15Q&A
related talks
28mLyra: Disrupting the full-text search industry with JS
19mDisrupting Full Text Search Industry With JavaScript
41mAn introduction to search-index.js
16mJavaScript Text Analytics for User Behavior Intelligence
32mWhat's Happening in the JavaScript Ecosystem?
48mJavaScript in Browser Performance
24m10 things I learned making the fastest JS server runtime in the world
43mThe Compute Revolution You’re Ignoring: JavaScript in Science