Wasn't meaning to sound angry, just shocked (at MA, not at you). Using MD5 for password hashing is crazy irresponsible, and this has been well known for around 25 years now.
As for your questions: MD5 always outputs 128 bits, and hashing functions don't have any maximum input.
MD5 was never used in crypto currency, btw. It's too fast and too narrow.
Ah it must been your forum picture that got me thinking your angry then, my bad.
Actually I didn't write MD5 had anything to do with crypto currencies.
But since you brought it up, RIPEMD-160 and SHA256 are used for bitcoin addresses.
SHA-2 is susceptible to length extension attacks and SHA256 is a weak variant of SHA-2 algorithm.
Original RIPEMD has collision weaknesses since 1996' so not entirely dissimilar situation to MD5.
But people still use them for the same reason as any variant of MD5, for the speed and size.
Regarding SHA-3 it has no known weaknesses unlike previous version SHA-2.
Although, heard some are still upset at how the SHA-3 competition went......
If concerned about SHA-3 (no known weaknesses), then suggest avoiding PBKDF2 which has known weaknesses.
Agree that bcrypt is a good one, but there's also Whirlpool, Blake2, and scrypt.
Anyways your reply to both MD5 lengths is correct.
In the past I found half the people will answer the input length question wrong.
And anyone can confirm that by decrypting my MD5 example in previous post.
Don't worry no salt added, just use one of the oldest crypto tricks in the book.
Have fun.