$body =~ s/<.+?>//g;

<! Das >-Zeichen sollte man besser escapen -->
# oder:
<input type="button" value="--->" name="Knopf">

package Sources::HtmlParser;

use strict;
use warnings;

use base 'HTML::Parser';

sub parseText
{
   my($self, $text) = @_;

   $self->{mytext} = "";
   $self->parse($text);
}


sub text
{
   my($self, $origtext, $is_cdata) = @_;

   $self->{mytext} .= $origtext;

   1;
}

sub get_text
{
   my($self) = @_;

   return $self->{mytext};
}

1;

my $hp = Sources::HtmlParser->new();
$hp->parseText($httpbody);
$httpbody = HTML::Entities::decode_entities($hp->get_text);
$httpbody =~ s/<!--.+?-->//gms;

use HTML::Parser;
my $parser = HTML::Parser->new(
    start_h => [ \&_starttag, 'self, tagname, attr' ],
    end_h   => [ \&_endtag,   'self, tagname' ],
    text_h  => [ \&_text,     'self, dtext' ]
  );

$parser->parse($email_text);

sub _starttag {
    my ($self, $tag, $attr) = @_;
    $self->{'_body'} = 1 if($tag eq 'body');
}

sub _endtag {
    my ($self, $tag) = @_;
    $self->{'_body' } = 0 if($tag eq 'body'  );
}

sub _text {
    my ($self, $dtext) = @_;

    $dtext =~ s/\A\s+//;
    $dtext =~ s/\s+\z//;
    return() unless ( length($dtext) > 0 and $dtext =~ /[^\s]/ );

    if ($self->{'_body'}    == 1) {
      print $dtext;
    }
}

#!/usr/bin/perl -w

use strict;
use warnings;
use HTML::FormatText;
use HTML::Parse;

my $html;
my $formatter;


$html = HTML::Parse::parse_htmlfile("beispiel.html");
# alternativ ginge auch... $html = HTML::Parse::parse_html($text);
# ...um den HTML-Text aus einer Variablen zu holen

$formatter = HTML::FormatText->new();

print $formatter->format($html);