"A couple of grand" can actually get you quite far when you're going at it with a start-up attitude. Assuming that you are willing to spend some time on this as well as some money, there are plenty of things you can do. There are already a lot of good answers here, so I'll try keep repetition to a minimum, but based on my experience working on UX at a start-up, here are some of my thoughts.

1. **Get to know your users**. Even if you don't have any yet. You are building a product, so you will need to do this anyway (how else will you know your market). Get specific. very few products are for "everybody". You can always identify different types of users that have different demographics, needs etc. Find out what makes each type tick (preferably by talking to them) and write this down (You can take this one step further and construct [persona](http://webdesign.tutsplus.com/articles/user-experience-articles/defining-and-applying-personas-to-ux-design/) if you like).
2. **Get early feedback**. You don't need to have a finished product to get feedback. You can get feedback on lots of things, from stories to crappy drawings on the back of a napkin. Again, lots of this stuff you will have anyway. You probably sketched some screens of your interface before you started coding HTML – show those to some people. Make sure you ask the right things to the right people: ideally you get feedback from real potential users, but if you can't get your hands on those, you can also ask friends or even colleagues (those are obviously biased, but can still get you feedback on whether screens are confusing, buttons are not labeled clearly etc.).
3. **Iterate**. Keep asking for feedback as you go along.
4. **Use the web**. If you are a web startup, you can get a lot of feedback from the web, some of it for free. Websites such as [FiveSecondTest](http://www.fivesecondtest.com) are great, because you can post a screenshot and get feedback from a total stranger for free. You can also look into things like Amazon Mechanical Turk and ask people for feedback on Twitter.
5. **Use Unmoderated Remote User Testing**. For less than 200 dollars/pounds/whatever you can get a few complete strangers to use your website, performing tasks you give them, while recording a video of their actions and thoughts. This has many limitations (they are not real users and may be too experienced, you cannot interact with them etc.), but the information you get from it has very good value for money.
6. **Get in touch with your first users**. Make it easy for them to get in touch with you, but also try to reach out to them when they talk about you on Twitter or something. You'd be surprised how willing to talk some people are and how much valuable information you can get from them. If your site is live, you can even recruit people from among users of your site using popups (but target them properly so you don't annoy and scare away users/customers. Services like ethnio can help).
8. **Measure**. Define what indicates different levels of "success" for your product. Measure things like registrations, logins, active users, purchases, everything. You can only detect problems if you know what a problem looks like.
9. **Use A/B testing**. Once you launch, this can help you test the impact of changes you make. Be careful that you're actually interpreting the results right, it's easy to draw the wrong conclusions. To your homework and define the right performance metrics and set up the test in a solid way.
10. **Get expert opinions**. If you can afford it, it may be worth to pay someone to give their expert opinion. I think this is most valuable for solving problems, rather than finding them. If you have identified a problem from user feedback or usability testing, an experienced designer might be able to help you understand why the problem exist and propose a solution.

A lot of this may seem like blasphemy for a UX professional with a budget, but doing anything is better than doing nothing at all.