Find and replace in bulk
Supports regex and capture groups ($1 references), check the match count before you replace.
with (\d+) capture groups like this, use $1 references;$& means the whole match. A wrong regex errors out immediately and never changes the text.
When to use it
- Exported data contains full-width brackets, double spaces and non-breaking spaces (U+00A0) that need clearing in one pass — a regex like [\u00A0 ]+ beats editing dozens of spots by hand.
- You want 2024-01-05 rewritten as 2024/01/05 everywhere, or a country code prepended to every phone number while keeping the digits via a capture group.
When not to
- You need multi-line matching or ^ and $ per line: the page always uses g plus an optional i, with no m or s, so ^ matches only the very start and . does not cross a newline.
- The replacement needs a decision (different cases map to different text, or a value must be computed): a script such as sed or Python is more reliable than stacking regex.
Replacement runs locally and the text is not uploaded. Limits: the regex always uses g plus an optional i, so no m, s or u — ^ and $ apply to the whole text, . does not cross lines and \p{...} is unavailable; the tool treats input as plain text and does not know about HTML or CSV structure; a pass over several MB visibly stutters.
Frequently asked
If my regex is wrong, will it damage the text?
No. The input box is never written back; results only appear in the output box. When the regex fails to compile the replacement never runs, and the output box keeps its previous content. The catch is that the error goes to the browser console and nothing appears on the page, so the button seems to do nothing — an unmatched bracket or an unclosed [ usually explains it.
Why does \p{L} or \u{1F600} match nothing?
The regex is built with new RegExp(..., 'gi'), without the u flag, so \p is read as a literal p and \u{1F600} as the literal text u{1F600}: no error, no match. For Unicode properties or code-point escapes, write the surrogate pair (\uD83D\uDE00) or the character itself.