Find and replace text

Find and replace is the job you do in an editor, until the text is in a browser because it came out of OCR, a PDF extract, or a CSV you do not want to open in Excel. imissfiles runs the search in this tab with a live preview: matches highlighted in the source, the result of replace-all in a second panel, and a count. Case sensitive, whole word, and regular-expression mode are checkboxes, not hidden menus. Nothing is uploaded, which matters when the haystack is a letter, a dump of emails, or a config with secrets.

Plain find is literal. Whole word wraps the query in a word boundary so “art” does not hit “party.” Case sensitive does what it says. Regex mode treats the find field as a JavaScript regular expression. If the pattern is invalid, you get the engine’s error inline — not a silent empty result. In regex mode the replace field understands capture groups: $1, $2, $&. A pattern like (\d{4})-(\d{2})-(\d{2}) with replace $2/$3/$1 is the usual date shuffle. Test on the example before you run it on a 4,000-line file.

Replace all is the nuclear option and the default people want. Step-through is for when you should not trust the pattern yet: Previous / Next moves the highlight, Replace this applies the substitution at the current match and keeps going. The stats line reads like “34 matches found · 34 would be replaced · viewing 4 of 34.” After replace-all, copy or download the output. The original box does not change until you apply, so you can still tweak the pattern.

Regex is powerful and easy to overfit. A greedy .* will eat more than you wanted. An unescaped dot matches any character. We do not try to sandbox your pattern beyond what the browser already does; we do run it in a worker so a catastrophic backtrack is less likely to freeze the UI, and we warn past about 5 MB. If the worker times out, shrink the haystack.

The natural workflow on this site is extract, then edit. PDF to Text, Image to Text, or OCR PDF first, then find-and-replace the OCR leftovers (“rn” that should be “m”, a wrong client name). Diff Checker is how you confirm the replacement did not take extra bites. Word Counter is how you confirm a caption still fits. All of it stays on this device.