@netgrep/netgrep

The main netgrep package. See the main README for more information.

netgrep streams a URL's response through ripgrep's real regex engine and answers the moment a matching line arrives — without waiting for the last byte, and without holding the file in memory. No index to build, no backend to run.

It runs in a browser and nowhere else: it needs fetch with a streaming response body, and it is ESM only. The file has to be one the browser can fetch anonymously — a cross-origin URL needs Access-Control-Allow-Origin, and netgrep puts nothing on the request, so anything behind a login is out of reach however permissive its CORS policy.

It answers one question per URL — does this pattern occur? — as a boolean, plus the first matching line if you pass capture: 'line', or that line with each match's position within it under capture: 'line-ranges'. That is the whole result: no ranking, no match counts, no line numbers, no positions in the file. It costs a ~1.17 MB WebAssembly download (~500 KB gzipped) once per page load, and it has known limitations worth reading first.