With RTNI_light one can integrate matrices symbolically with respect to Haar unitary matrices of size n.
RTNI_light is a light and user-friendly version of the program called RTNI, originally developped in the following paper:
M. Fukuda, R. Koenig, I. Nechita "RTNI - A symbolic integrator for Haar-random tensor networks",
arXiv:1902.08539 [quant-ph].
The programs of the above paper can be found at the GitHub page:
RTNI at GitHub.
The web-version of RTNI_light can be found at the pythonanywhere website:
RTNI_light web
but there are limitations to save computational resource of the server.
Manual can be found here:
Manual.