Hi, thanos1983. I'd like to point out, that use of non-latin1 characters in Perl source, without use utf8; is misleading. Perhaps your example might be modified to this (note the differences, while output is deceptively the same):

#!/usr/bin/perl
use strict;
use warnings;
use URI::Escape;
use feature 'say';

use utf8;
use Encode qw/ encode decode /;
binmode STDOUT, ':utf8';

my $str = "Character one: ω character two: ∞";
my $hex_code = uri_escape_utf8( $str );
say $hex_code;

my $string = decode 'UTF-8', uri_unescape( $hex_code );
say $string;

__END__

To answer Anonymous Monk's direct question:

use strict;
use warnings;
use PDF::API2;
use URI::Escape;
use Encode qw/ decode /;

my $percent_encoded_str = '%CF%89%20%E2%88%9E';
my $octets              = uri_unescape $percent_encoded_str;
my $proper_unicode_str  = decode 'UTF-8', $octets;

my $pdf  = PDF::API2-> new;
my $page = $pdf-> page;
my $text = $page-> text;

my $ttf_font = $pdf-> ttfont( 'DejaVuSans.ttf' );

$text-> font( $ttf_font, 20 );
$text-> translate( 50, 700 );
$text-> text( $proper_unicode_str );

$pdf-> saveas( 'test.pdf' );
[download]