Instead of a search pattern, give jgrep a plain-English description like "a user is getting frustrated", and it keeps the lines that match in meaning. It works on files or on a live stream, such as an app log.
Each line is judged by Jev in about a fifth of a second for a thousandth of a cent, and it can show the match score next to each line so you can sort by it.
What you get
- Full Python source code, MIT licensed
- Command-line tool with benchmarks and docs






