Either #1 or #2 sound like they could be a good approach for what you want to accomplish. Which choice is ultimately better will depend on different factors, but hopefully I can help in that regard. (I'm assuming your early adoption phase is mostly for product quality, as opposed to marketing—that line gets blurred so often with web applications.)
Direct, Hands-On Approach
The challenge with #1 is that the amount of effort required isn't necessarily scalable. If your SAAS offering doesn't face a lot of competition and also solves a major problem for people, you might get a high response rate to your pitches. But what happens if you don't get that high response rate?
When it comes to beta and beta-like testing phases, people often make the assumption that if they have 20 people who sign up for the test, they'll get 20 people (or at least close to it) offering feedback. Unfortunately, that's not realistic. Participation rates can often be as low as 1/4 or 1/3. Since it sounds like you want to put a lot of care into recruiting and managing this process, you might see much better participation along the lines of say 1/2 or 2/3. But you have to factor these things in when picking your numbers. If you've concluded that getting feedback from 20 users is important to the goals of the user test, then you probably want to set your goal for the number of users you let into this test phase at 40. If only 10% of the people you pitch express interest in the project, then you're looking at 400 people you need to contact.
Your actual numbers may be much better, which could really swing things in favor of using this approach. But the basic idea I'm trying to convey is that you'll probably get better results through direct contact, but much more effort might be required.
Online Invite Request
This approach works just fine. In fact, when the company I work for is managing beta tests for other software and hardware companies, this is usually how we recruit for their test communities. But it does require you to figure out how you're driving traffic to the sign-up page and how to optimize its conversion rate.
Since you do get the luxury of screening when you're doing online sign-ups, there are two simple questions you can ask that will really do a lot of the work for you. (1) Why do you want to test this product? and (2) What makes you a great tester for this product? If someone isn't willing to answer those questions, it's probably a good indicator that they wouldn't have participated anyway. And people who give you high-quality answers will probably give you high-quality feedback once they're using your application.
Finally, while you're only looking for a small number of testers, I'd still suggest harvesting sign-ups as if you were looking for many more. If you decide to run a full beta later, you'll already have a list of candidates. And it's also advantageous from a marketing perspective. So, by going this route, you could find you have more upside in the long term, but less of a relationship established with your testers.
Expectation management is important, but it will be clear enough either way you decide to handle inviting your users. It's better to focus on your goals for the test and which approach will yield the most benefit to you (and, in turn, your future customers). Definitely nurture and try to foster a sense of community within the early adoption test. Sounds like you're on the right track. Good luck!