Piping curl into a pager

E.g.
$ curl www.google.com | sed '/<[^>]*>/d' | less

Could this be used as the basis for a text-based browser? Curl has POST and cookies support too, and Guile Scheme could be used instead of javascript.

>>12,13
Next in plan is creating a 'lshttp' script (similar to ls) for listing all linked hyperlinks inside a given web page.
Something like
#!/bin/bash
curl $1 | sed '/<\s*a \s*href\s*=\s*"([^"])"/gi!d'
(Haven't tested yet.)