The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.


GraphQL::Client - A GraphQL client


version 0.605


    my $graphql = GraphQL::Client->new(url => 'http://localhost:4000/graphql');

    # Example: Hello world!

    my $response = $graphql->execute('{hello}');

    # Example: Kitchen sink

    my $query = q[
        query GetHuman {
            human(id: $human_id) {
    my $variables = {
        human_id => 1000,
    my $operation_name = 'GetHuman';
    my $transport_options = {
        headers => {
            authorization => 'Bearer s3cr3t',
    my $response = $graphql->execute($query, $variables, $operation_name, $transport_options);

    # Example: Asynchronous with Mojo::UserAgent (promisify requires Future::Mojo)

    my $ua = Mojo::UserAgent->new;
    my $graphql = GraphQL::Client->new(ua => $ua, url => 'http://localhost:4000/graphql');

    my $future = $graphql->execute('{hello}');

    $future->promisify->then(sub {
        my $response = shift;


GraphQL::Client provides a simple way to execute GraphQL queries and mutations on a server.

This module is the programmatic interface. There is also a "CLI program".

GraphQL servers are usually served over HTTP. The provided transport, GraphQL::Client::http, lets you plug in your own user agent, so this client works naturally with HTTP::Tiny, Mojo::UserAgent, and more. You can also use HTTP::AnyUA middleware.



The URL of a GraphQL endpoint, e.g. "http://myapiserver/graphql".


Whether or not to "unpack" the response, which enables a different style for error-handling.

Default is 0.



The package name of a transport.

This is optional if the correct transport can be correctly determined from the "url".


The transport object.

By default this is automatically constructed based on "transport_class" or "url".



    $graphql = GraphQL::Client->new(%attributes);

Construct a new client.


    $response = $graphql->execute($query);
    $response = $graphql->execute($query, \%variables);
    $response = $graphql->execute($query, \%variables, $operation_name);
    $response = $graphql->execute($query, \%variables, $operation_name, \%transport_options);
    $response = $graphql->execute($query, \%variables, \%transport_options);

Execute a request on a GraphQL server, and get a response.

By default, the response will either be a hashref with the following structure or a Future that resolves to such a hashref, depending on the transport and how it is configured.

        data   => {
            field1  => {...}, # or [...]
        errors => [
            { message => 'some error message blah blah blah' },

Note: Setting the "unpack" attribute affects the response shape.


There are two different styles for handling errors.

If "unpack" is 0 (off, the default), every response -- whether success or failure -- is enveloped like this:

        data   => {...},
        errors => [...],

where data might be missing or undef if errors occurred (though not necessarily) and errors will be missing if the response completed without error.

It is up to you to check for errors in the response, so your code might look like this:

    my $response = $graphql->execute(...);
    if (my $errors = $response->{errors}) {
        # handle $errors
    else {
        my $data = $response->{data};
        # do something with $data

If unpack is 1 (on), then "execute" will return just the data if there were no errors, otherwise it will throw an exception. So your code would instead look like this:

    my $data = eval { $graphql->execute(...) };
    if (my $error = $@) {
        my $resp = $error->{response};
        # handle errors
    else {
        # do something with $data

Or if you want to handle errors in a different stack frame, your code is simply this:

    my $data = $graphql->execute(...);
    # do something with $data

Both styles map to Future responses intuitively. If unpack is 0, the response always resolves to the envelope structure. If unpack is 1, successful responses will resolve to just the data and errors will fail/reject.



Please report any bugs or feature requests on the bugtracker website

When submitting a bug or request, please include a test-file or a patch to an existing test-file that illustrates the bug or desired feature.


Charles McGarvey <>


This software is copyright (c) 2020 by Charles McGarvey.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.