Capable GraphQL client

graphql, http, async, aiohttp, api-client, graphql-client, http-client, requests
pip install quiz==0.2.2



Build status Test coverage Dependabot Documentation status
Development status Latest version License Supported python versions Black code style

Capable GraphQL client for Python.


  • Sync/async compatible, pluggable HTTP clients.
  • Auto-generate typed and documented python APIs
  • ORM-like syntax to write GraphQL.

Note that this project is in an early alpha stage. Some features are not yet implemented (see the roadmap below), and it may be a little rough around the edges. If you encounter a problem or have a feature request, don't hesitate to open an issue in the issue tracker.


A quick 'n dirty request to GitHub's new V4 API:

>>> import quiz
>>> query = '''
...   {
...     repository(owner: "octocat", name: "Hello-World") {
...       createdAt
...       description
...     }
...   }
... '''
>>> quiz.execute(query, url='',
...              auth=('me', 'password'))
{"repository": ...}


  1. Adaptability. Built on top of snug, quiz supports different HTTP clients

    import requests
    result = quiz.execute(query, ..., client=requests.Session())

    as well as async execution (optionally with aiohttp):

    result = await quiz.execute_async(query, ...)
  2. Typing. Convert a GraphQL schema into documented python classes:

    >>> schema = quiz.Schema.from_url('',
    ...                               auth=('me', 'password'))
    >>> help(schema.Repository)
    class Repository(Node, ProjectOwner, Subscribable, Starrable,
     UniformResourceLocatable, RepositoryInfo, quiz.types.Object)
     |  A repository contains the content for a project.
     |  Method resolution order:
     |      ...
     |  Data descriptors defined here:
     |  assignableUsers
     |      : UserConnection
     |      A list of users that can be assigned to issues in this repo
     |  codeOfConduct
     |      : CodeOfConduct or None
     |      Returns the code of conduct for this repository
  3. GraphQL "ORM". Write queries as you would with an ORM:

    >>> _ = quiz.SELECTOR
    >>> query = schema.query[
    ...     _
    ...     .repository(owner='octocat', name='Hello-World')[
    ...         _
    ...         .createdAt
    ...         .description
    ...     ]
    ... ]
    >>> str(query)
    query {
      repository(owner: "octocat", name: "Hello-World") {
  4. Offline query validation. Use the schema to catch errors quickly:

    >>> schema.query[
    ...     _
    ...     .repository(owner='octocat', name='Hello-World')[
    ...         _
    ...         .createdAt
    ...         .foo
    ...         .description
    ...     ]
    ... ]
    SelectionError: SelectionError on "Query" at path "repository":
        SelectionError: SelectionError on "Repository" at path "foo":
            NoSuchField: field does not exist
  5. Deserialization into python objects. Responses are loaded into the schema's types. Use . to access fields:

    >>> r = quiz.execute(query, ...)
    >>> r.repository.description
    "My first repository on GitHub!"
    >>> isinstance(r.repository, schema.Repository)

    If you prefer the raw JSON response, you can always do:

    >>> quiz.execute(str(query), ...)
    {"repository": ...}


quiz and its dependencies are pure python. Installation is easy as:

pip install quiz


After you've cloned the repo locally, set up the development environment with:

make init

For quick test runs, run:


To run all tests and checks on various python versions, run:

make test

Generate the docs with:

make docs

Pull requests welcome!

Preliminary roadmap

Feature status
Input objects v0.3.0
better query validation errors v0.3.0
more examples in docs v0.3.0
executing selection sets directly v0.3.0
introspection fields (i.e. __typename) v0.3.0
custom scalars for existing types (e.g. datetime) v0.3.0
improve Object/Interface API v0.3.0
value object docs v0.3.0
Mutations & subscriptions v0.3.0
Inline fragments v0.3.0
Fragments and fragment spreads v0.4.0
py2 unicode robustness v0.4.0
Mixing in raw GraphQL planned
Module autogeneration planned
Type inference (e.g. enum values) planned
Variables planned
Directives planned
Integer 32-bit limit planned
converting names from camelcase to snake-case idea
Autogenerate module .rst from schema idea
Autogenerate module .py from schema idea
Escaping python keywords idea
Handling markdown in descriptions idea
Warnings when using deprecated fields idea
Handle optional types descriptions in schema idea
Returning multiple validation errors at the same time idea
Explicit ordering idea