I think there's a lot of room for improvement, but I'll say up front that as a prototype, this is very good. :-)
I did two playthroughs of the game, with a bot opponent; not ideal, I know, but I'm home alone at the moment and wanted to see what it was like. Some things I noticed;
* It would be nice to have the letters guessed/unguessed visible even when not typing. Knowing which consonants are already gone is a huge help in determining whether you want to guess a consonant at all, or do something else. It's also a help when trying to work out the answer before committing to solving ("the second word could be MINERAL...wait, did we already guess E?").
* There's a weird disincentive to use the second action on your turn. If you have four energy, and all of your cards cost two, then you'll, say, guess a consonant for your first action. Then, you could guess another consonant and get maybe a couple hundred dollars--but it's giving a lot of information, and you won't have enough energy left to use that information!
That's especially true of "get the first letters". When the bot used it, it lit up the grid with *S??I???????, I??I??IS*, which was easily enough to get ILLINOIS and, from that, SPRINGFIELD. (There may have been another letter or two already revealed.) So that's almost never going to be your best play, since it uses all your energy and probably gives away the answer for the next person. (Unless you got lucky enough to get the bonus energy.)
It's a tough balance, and I think it's one of the reasons the Wheel-of-Fortune model actually works: there's always incentive to guess letters, because guessing letters can give you more money than just solving; and if you guess correctly, your turn doesn't end and you can spin to guess more. (I almost wonder whether the right strategy is to never guess any letters, and just wait for your opponents to fill the board enough for you to guess the answer. Of course, if everyone in the game does that...)
* I think there's opportunity for one-player mode, in which you lose money for making wrong guesses (maybe all guesses cost money?), and you try to maximize your money.
All of this is, of course, if you want to keep working on it. If not, you can rest happy, knowing you made a very good prototype. :-)