Regex Tester (live matches, groups, replace)

What flags do

g (global): find all matches, not just the first. Required for replace-all behavior. i (case-insensitive): /[a-z]/i matches both cases. m (multiline): ^ and $ match at line breaks instead of only string start/end. s (dotall): . matches newlines as well. u (unicode): proper handling of code points above 0xFFFF (emoji, CJK in surrogate pairs).

Most people start with /g/i/m. /s/ becomes useful for multi-line text matching. /u/ is recommended whenever you might encounter emoji or non-BMP characters — without it, /./ won't match a 4-byte emoji as a single character.

Capture groups and replace

Parentheses create a capture group: /(\d{4})-(\d{2})/ on '2026-05' captures '2026' and '05' separately. Reference them as $1, $2 in the replacement. Named groups use (?<name>...) and reference $<name> in replace.

Common replace tricks: $& inserts the full match. $1, $2 etc. insert capture groups. $$ inserts a literal $. The replace preview here lets you check before running the same pattern on a real text — much safer than running it directly on production data.

Pitfalls to watch for

Backslash escaping. To match a literal '.', use '\.'. In a JavaScript source string this becomes '\\.' but in this tool you type just '\.' since it's the regex source.

Catastrophic backtracking. Patterns like /(a+)+/ on 'aaaa…!' can take forever. If your test seems frozen, your regex may have nested quantifiers. Simplify or use possessive/atomic groups (not available in JS regex; use careful design instead).

Greedy vs lazy. /<.+>/ on '<a><b>' matches '<a><b>' (greedy). /<.+?>/ matches '<a>' (lazy). Pick the right one for your use case.