// the find
jameslittle230/stork
🔎 Impossibly fast web search, made for static sites.
Stork is a Rust/WASM static-site search tool: you run a CLI to build a binary index file from your content, then drop a JS snippet on your page that searches it client-side with no server round-trips. It's aimed at Jamstack blogs and docs sites that don't want to pay for Algolia or run Elasticsearch.
The architecture is genuinely smart — the index is a single binary file you upload once, and all searching happens in WASM in the browser with no backend. Result highlighting works at the excerpt level, showing exact match context rather than just page titles. The Rust implementation means indexing is fast even for large content sets. TypeScript for the browser JS layer and tests for both halves is more disciplined than most side projects.
The README opens with the maintainer announcing he's winding down the project — this is effectively abandoned as of 2023, which rules it out for any production use. The index format is TOML-configured per-file, meaning you need a build step that knows your full content structure; there's no crawler or plugin ecosystem to generate configs automatically for common SSGs beyond a handful. No fuzzy matching — prefix search only, so typos return nothing. The CDN hosting (`files.stork-search.net`) referenced throughout the docs is presumably going to disappear eventually, which would silently break any site still using it.