Filter lines by text or pattern and keep or drop what matches, then number them or wrap each one in a prefix and suffix. Line numbers can be stripped again.
Waiting for text
Filter lines, number them, add a prefix or suffix. Nothing is uploaded.
The filtered lines appear here
Matching lines kept or dropped, optionally numbered and wrapped in a prefix and suffix.
Text Filter runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
Grep, nl and a trivial sed expression, in a box, for the many occasions when a log file is in a browser tab and a terminal is not in front of it. Filter the lines you want, number them, wrap each one in something.
Keep or drop is one switch rather than two modes. Lines containing a word and lines not containing it are the same question asked from either side, and every filter tool that offers only one of them sends you away to rewrite the pattern as a negation. Seven tests are available, including blank and duplicate, which test the line itself and so need no pattern — and the pattern field disappears for those rather than sitting there being ignored.
A pattern that does not compile is reported and the text is left alone, which is the right behaviour because a half-typed regular expression is the normal state while someone is writing one. Dropping every line in response would make the tool unusable for the thing it is for.
Numbering is reversible, and that is more useful than it sounds. Numbered output pasted from somewhere else — a code listing from a book, a stack trace from a web console, a document with line references — is the single most common thing that needs cleaning before it can be used. The strip is deliberately conservative: it removes a leading number followed by a separator, and leaves a line alone when what follows the number does not look like one, because a line that legitimately begins with a figure is far more common than a mis-numbered one.
The prefix and suffix fields turn a list of values into a list of quoted strings, a set of bullet points, or a column of SQL inserts, which is the kind of five-second job that otherwise takes a multi-cursor editor and two minutes. Nothing is uploaded.
Yes, with the keep-or-drop switch. There is no need to rewrite your pattern as a negation, which is what a tool offering only one direction forces you to do.
The error is reported and the text is left alone. A half-typed pattern is normal while you are writing one, so dropping every line in response would make the tool useless for its purpose.
Because the blank and duplicate tests examine the line itself and have nothing to match against. A field that is present but ignored is worse than one that is not there.
It removes a leading number followed by a separator. When what follows the number does not look like a separator, the line is left alone, because a line that genuinely begins with a figure is much more common than a mis-numbered one.
The test identifies a line as a duplicate when an identical line appeared earlier, so the first occurrence is not a duplicate. Combine it with the keep-or-drop switch to keep either the first instances or only the repeats.
No. Filtering happens in your browser and nothing reaches a server or analytics, which is what makes it safe for a log with real data in it.