It's a web page that lets you create capture filters that match strings in TCP payloads.
It takes the string you enter, splits it into 1, 2, or 4 byte chunks, converts them to numbers, and creates a capture filter that matches those numbers at the offset you provide.
It should handle most UTF-8 characters but this hasn't been tested.
You can use it to filter things like top-level HTTP requests ("GET / HTTP/1."), HTTP responses ("HTTP/1."), POP3 logins ("USER"), and lots of other things.
Matching strings at arbitrary locations. You can't do that with capture filters (BPF doesn't support it) You need to use the "matches" or "contains" display filter operators instead. You'll have to use the "matches" display filter operator for case insensitive matching as well.
It makes sure we skip over any TCP options that might be present. See Sake's explanation for more details.