TXR contains an original Lisp implementation (TXR Lisp), which can be used on its own for tasks that don't require the TXR pattern-based extraction language. Files with a .tl suffix are processed by the txr executable as TXR Lisp.

There is an interactive Lisp REPL (just run txr with no arguments, or txr -i to suppress the banner message). It has history, completion, multi-line editing with cut and paste and all those goodies. You can interrupt long computations with Ctrl-C, which actually works by turning a signal into an exception caught by the REPL.

TXR is ported to Windows; you can download a nicely wrapped installer version, your choice of 32 or 64 bit.

https://bintray.com/kazinator/Binaries/TXR/

The REPL works nicely on Windows, in its own console window or right out of cmd.exe.

TXR has a built-in mechanism for wrapping up TXR code as an executable application (which is geared toward Windows, but works on other platforms):

http://www.nongnu.org/txr/txr-manpage.html#N-037ACEFC

On Windows, use the txr-win.exe flavor of the binary to avoid opening a console window.

TXR example combining OOP with delimited continuations to implement "synchronous concurrency":

https://rosettacode.org/wiki/Synchronous_concurrency#TXR

There is more to TXR than "unusual notation for matching documents and extracting data". The @ character occurs only once in this example. :)