I think I need to choose between:
1. progressing fast now, but slowing down when - and if - the site takes off, as I will lose some flexibility I'd have with Rails.
2. progressing slower now and go rails (or even django)
I think I'll go with 1.
You won't get any customers until it's at least out of the door. Seriously, JFDI if you can and get it out. Worrying about flexibility if you're successful is a luxury problem, and if you're not successful then you just saved a load of time.