Skip to content


Repository files navigation


Quality Gate Status Total alerts Language grade: Python

Tartiflette is a GraphQL Server implementation built with Python 3.7+.



Read this blogpost about our motivations TL; DR We reached the limits of Graphene, we wanted to build something which met certain requirements:

  • Offers a better developer experience that respects the Python mindset
  • Uses SDL (Schema Definition Language)
  • Uses asyncio as the sole execution engine
  • Be 100% open source


The first milestone is behind us, we are now on the road to the milestone 2.


Discover Tartiflette with our fabulous tutorial on


import asyncio

from tartiflette import Resolver, create_engine

async def resolver_hello(parent, args, ctx, info):
    return "hello " + args["name"]

async def run():
    engine = await create_engine(
        type Query {
            hello(name: String): String

    result = await engine.execute(
        query='query { hello(name: "Chuck") }'

    # {'data': {'hello': 'hello Chuck'}}

if __name__ == "__main__":

More details on the API Documentation


Tartiflette is available on

While the project depends on libgraphqlparser, wheels are provided since version 1.4.0, ensuring that no system dependency is required.

To install the library:

pip install tartiflette

Building from source

If you use a platform incompatible with the provided wheels, you'll need to install cmake to build libgraphqlparser in order to install the library.


brew install cmake


apt-get install cmake

HTTP server implementations

tartiflette library itself is transport agnostic, but to simplify integration with existing HTTP servers, two different libraries are available:


How to contribute to the documentation?

As you may know, the documentation is hosted on This fabulous website is built thanks to another amazing tool, docusaurus.

The content of the documentation is hosted in this repository, to be as close as possible to the code. You will find everything you need/want in the folder /docs.

How to run the website locally?

We built a docker image for the documentation (tartiflette/ on docker hub), which allow us to provide you an easy way to launch the documentation locally, without installing a specific version of node.


  • Docker
  • Docker Compose
  • Make
make run-docs

Every change you will make in the /docs folder will be automatically hot reloaded. 🎉