Filter / grep a log file

That depends on your device, and this is the tool most likely to hit the limit. The whole file is read into memory as text, split into lines and filtered there, with the kept lines held as a second copy. A log genuinely in the multiple gigabyte range can exceed what a browser tab holds. It does not stream from disk, so for those, use grep on a machine that can.

Plain substring by default. Turn useRegex on and the pattern is compiled as a JavaScript regular expression, and an invalid one is refused up front with "invalid regex" and the reason. caseSensitive is on by default and applies to both modes; turning it off lowercases both sides for the substring search and adds the i flag for the regex.

Yes. Turn invert on and every line that does not match is kept, which is the equivalent of grep -v. It combines with the other options, so you can invert a case insensitive regex. The count line still reports how many lines were kept out of the total, counted after the inversion, and no file is written when nothing is kept.

Your file name with -filtered inserted before the extension, keeping the extension you had, or .log if the file had none. So app.log becomes app-filtered.log. Kept lines are joined with LF and a trailing newline is added, so CRLF input comes back as LF.

Because nothing matched. When zero lines match, the tool reports "0 of N lines matched" and writes no file, since an empty download would be misleading. Whenever at least one line matches you get both the count and the file.

No. The file is read and filtered in a Web Worker in your own tab, with no upload and no account. That matters for logs in particular, since they often carry addresses, tokens and user data you do not want on a server you do not control.