Parallel::Downloader - simply download multiple files at once


version 0.132071


    use HTTP::Request::Common qw( GET POST );
    use Parallel::Downloader 'async_download';

    # simple example
    my @requests = map GET( "" ), ( 1..15 );
    my @responses = async_download( requests => \@requests );

    # complex example
    my @complex_reqs = ( ( map POST( "", [ type_id => $_ ] ), ( 1..60 ) ),
                       ( map POST( "", [ type_id => $_ ] ), ( 1..60 ) ) );

    my $downloader = Parallel::Downloader->new(
        requests => \@complex_reqs,
        workers => 50,
        conns_per_host => 12,
        aehttp_args => {
            timeout => 30,
            on_prepare => sub {
                print "download started ($AnyEvent::HTTP::ACTIVE / $AnyEvent::HTTP::MAX_PER_HOST)\n"
        debug => 1,
        logger => sub {
            my ( $downloader, $message ) = @_;
            print "downloader sez [$message->{type}]: $message->{msg}\n";
    my @complex_responses = $downloader->run;


This is not a library to build a parallel downloader on top of. It is a downloading client build on top of AnyEvent::HTTP.

Its goal is not to be better, faster, or smaller than anything else. Its goal is to provide the user with a single function they can call with a bunch of HTTP requests and which gives them the responses for them with as little fuss as possible and most importantly, without downloading them in sequence.

It handles the busywork of grouping requests by hosts and limiting the amount of simultaneous requests per host, separate from capping the amount of overall connections. This allows the user to maximize their own connection without abusing remote hosts.

Of course, there are facilities to customize the exact limits employed and to add logging and such; but async_download is the premier piece of API and should be enough for most uses.



Can be requested to be exported, will instantiate a Parallel::Downloader object with the given parameters, run it and return the results. Its parameters are as follows:

requests (required)

Reference to an array of HTTP::Request objects, all of which will be downloaded.


A reference to a hash containing arguments that will be passed to AnyEvent::HTTP::http_request.

Default is an empty hashref.


Sets the number of connections allowed per host by changing the corresponding AnyEvent::HTTP package variable.

Default is '4'.


A boolean that determines whether logging operations are a NOP or actually run. Set to any true value to activate the logging.

Default is '0'.


A reference to a sub that will receive a hash containing logging information. Whether that sub then prints them to screen or into a database or other targets is up to the user.

Default is a sub that prints to the screen.


The amount of workers to be used for downloading. Useful for controlling the global amount of connections your machine will try to establish.

Default is '10'.


A reference to a sub that will be called on completion of a request to build the response variable that will be returned for this request. It receives as parameters the body of the response, a hash ref of the response headers and the original request.

Default is a sub that returns the parameters wrapped in an array reference.


A boolean that determines whether the returned responses are sorted in the same order as the input requests. Can be useful to disable if build_response was overridden to not return an array or not return the request as the third element of the response array.

Default is '1'.



Runs the downloads for the given parameters and returns an array of array references, each containing the decoded contents, the headers and the HTTP::Request object.


Bugs / Feature Requests

Please report any bugs or feature requests through the issue tracker at You will be notified automatically of any progress on your issue.

Source Code

This is open source software. The code repository is available for public review and contribution under the terms of the license.

  git clone


Christian Walde <>


Christian Walde has dedicated the work to the Commons by waiving all of his or her rights to the work worldwide under copyright law and all related or neighboring legal rights he or she had in the work, to the extent allowable by law.

Works under CC0 do not require attribution. When citing the work, you should not imply endorsement by the author.