Regex tester

Building a pattern one piece at a time

The fastest way to a working regular expression is embarrassingly unglamorous: start by typing the literal text you want to find, watch it highlight, and only then generalise it step by step — swap the digits for a class like [0-9]+, allow optional parts with a question mark, anchor it with ^ or $ once the middle behaves. Because matches update live as you type, every small change gives you instant feedback, and a mistake shows up the moment you make it instead of three edits later. Just as important is what you paste in the text box: include lines the pattern should not match, not only the ones it should. A pattern tested purely against happy examples has a habit of matching half the document the moment it meets real data.

Capture groups do the extracting

Wrap part of a pattern in parentheses and it becomes a capture group — the piece of each match you actually want to keep. A pattern like ([0-9]{2}):([0-9]{2}) run against a page of text does not just find times; the table below the text lists each match with group one holding the hours and group two the minutes, plus the position where it was found. Add the g flag and you collect every occurrence in the text instead of stopping at the first. This is the honest core of most everyday regex work: not clever validation, but pulling structured values — dates, order numbers, prices — out of text that was never meant to be structured.

When a pattern hangs — and the safety net

Some patterns are traps. Nest one open-ended quantifier inside another — the classic shape is (a+)+ — and on the wrong input the engine tries an exponential number of ways to split the text before giving up. On most sites that freezes the whole tab; here the matching runs off in a background worker that is stopped after two seconds, so you get an explanation instead of a dead page. The cure is almost always to make the pattern more specific, so the engine has fewer ways to carve up the same text. One honest caveat: this tester uses JavaScript's engine, which is ideal if the pattern is destined for a script or webpage, but Python and PCRE differ on a few constructs — test in the final language too. When the values you extract are destined for a link or a query string, the URL encoder is the natural next stop.

How this tester works

Your pattern runs against the text using JavaScript's own regular expression engine — the same one your code will actually use if this is for a script or webpage. Matches are highlighted in the text and listed with their capture groups below.