There are already well-detailed youtube videos of installation for each Operating Systems in their documentary page.
If you come across an error, you can always ask here for a possible fix or workaround but you must indicate the specs and OS of the machine you're working on.

About the actual bruteforce, your token file must be created specifically for your case.
The positions of the known characters and possible characters must be indicated accurately and in a way that it'll create as little search space as possible.
Refer to this doc: btcrecover.readthedocs.io/en/latest/tokenlist_file/

"Seem" doesn't look promising since it has to be as accurate as possible.
You must also know the number of characters, how many are missing and uncertain.
This is so you can get a heads-up if it's possible or not before committing to purchase GPU/PC and set everything up.

BTW, BTCRecover's GPU support is still experimental so you might not get the performance boost that you're expecting with a GPU.