>>71720431honest answer:
Personally, understanding vtubers has a very steep learning curve because it's kind of like listening to computer nerds or car guys who don't explain anything so there are a lot of layers of context that you have to slowly process before you get to properly understand what they're saying a lot of the time. Not trying to discourage you though--just being realistic about the scope you're asking about.
If you're starting from scratch, learn the basics (hiragana, katakana, ha/ga, etc.).
Then install the japanese keyboard option (as opposed to the default english keyboard) and use
jisho.org or Google Translate to handwrite the kanji you want to know about.
Once you get better, you can pick up common grammar structures and vocabulary (for example, "possible consequences", negation, combining nouns, combining verbs, combining adjectives, and more). This will serve as your bread and butter and will help you make guesses as to what vtubers are saying. If you guess wrong, then you're learning from your mistakes. If you guess right, then well done! You found a good reference point that you can use as a youtube timestamp if you ever feel the need to check again.
Look up JP clippers and animators, seriously! They help a lot because vtubers often speak very fast and you'll need as many subtitles as you can get.