Regexer is a single-purpose site: a JavaScript regular-expression tester that shows live matches with per-group highlighting, previews String.replace() output with real $1/$<name> semantics, explains your pattern back to you in plain English, and warns when it spots a catastrophic-backtracking shape.

Why this site exists

The incumbents in this space — regex101, regexr — are excellent but share two gaps. First, they upload your pattern (and often your test text) for their sharing features; pasting real log lines or user data into them is a quiet privacy leak nobody thinks about. Second, they show you what matched but rarely why the pattern works the way it does. Regexer was built around those gaps: fully offline (the page is a static file; nothing you type can leave the tab) and explain-first (every token gets a plain-English line, including how the flags change ^, $ and .).

The third thing it does differently: honesty about regex itself. The pattern library ships the patterns everyone copy-pastes — email, URL, phone, date, IPv4, hex color — each with a written note about exactly where it lies to you. And the backtracking audit flags the (a+)+ family of bombs with a warning that says “heuristic, not proof” instead of pretending to be a formal analyzer. Articles like why email regex is a trap and catastrophic backtracking explained go deeper on both.

How it works

Everything runs in your browser’s own RegExp engine — the same engine that runs your production JavaScript, quirks included (\w is ASCII-only, . skips newlines without s, no g means one match). Group positions come from the d (hasIndices) flag added to an internal second regex, so indices shown are the engine’s own, not a re-implementation’s guess. The explainer is a hand-written tokenizer over the pattern text, and the backtracking audit is a structural heuristic over group spans — both are documented in the copy where they make trade-offs.

Who maintains it

Regexer is maintained by a small independent team that builds focused utility sites. The bar we hold it to: what the tool shows should match what new RegExp() does in your browser, and where a feature is heuristic or approximate the UI should say so. Found a divergence — a match reported wrong, an explanation that misreads a token, a warning that shouldn’t fire? Contact us — correctness reports get answered first.