Skip to main content

Django Trello Webhooks - Trello callback integration for Django.

Project description

Django application for managing Trello webhooks.


UPDATE 01-Dec-2014

The application can now be deployed directly to Heroku using their Deploy button. In addition, the test app that is bundled into this repo has been deployed and is live.

I am going to post further docs on the Trello board itself, so that will become the primary source for updates:


The app is now working, and deployable to Heroku (see below). The main outstanding issue is writing some proper docs, but until / unless there is any genuine external interest shown in this project I won’t be spending any formal time on it, so please don’t expect much more.


Trello webhooks can only be managed programmatically via the API, which is clever, but also a massive PITA. In addition, whilst creating a new webhook, Trello will immediately call the registered callback URL with a HEAD request to verify that it exists. This makes setting up webhooks fiddly and complex (esp. when experimenting in a local development environment).

This project is a Django app that makes managing Trello webhook easy. Well, easier. It exposes Trello webhooks through the Django admin site as standard Django models that can be created and deleted through a web UI. Incoming webhook events are then exposed via Django signals that are raised, and can be integrated into your application.

How does it work?

Each webhook is modelled as a Django model (Webhook), and each callback received on the webhook is modelled as a CallbackEvent object that contains the JSON payload (so it will maintain a complete history of all callbacks).

The method is used to register the webhook with the Trello API. The corresponding delete method is used to delete the webhook from Trello.

The app contains a single view function, api_callback, which receives the callback from Trello, and which also supports the synchronous activation of new webhooks by Trello. (When you create a webhook via the Trello API, they will immediately issue a HEAD request to the callback url supplied, so you need to be able to handle this immediately.)

The important bit is then how you use the callback in your application. This is done via Django signals. On each webhook callback, the app sends the callback_received signal, passing in the data received via the callback.

Your application then connects via this signal; below is taken from the included test_app, which sends the formatted event to HipChat:

from django.conf import settings
from django.dispatch import receiver

from test_app.hipchat import send_to_hipchat

from trello_webhooks.signals import callback_received

@receiver(callback_received, dispatch_uid="callback_received")
def on_callback_received(sender, **kwargs):
if settings.HIPCHAT_ENABLED:
    event = kwargs.pop('event')

If you wanted to filter out only certain events for sending:

def on_callback_received(sender, **kwargs):
    event = kwargs.pop('event')
    if event.event_type == 'commentCard':

There is a Django management command which can be used to synchronise any existing webhooks (in both directions), called sync_webhooks. Run on its own, without any arguments, this will look up all the webhooks in the local Django database, and sync then with Trello (creating them if they don’t already exist). It will also check Trello for any webhooks that it has registered that do not exist locally, and create them.

Rendering the payload

Once you’ve received a callback, along with its JSON payload, the next question is how to use it effectively. It is assumed (by me) that the core use case for this project is to pipe the events elsewhere - in our case it’s to HipChat, but other messaging services are available - you’ve could even go old-school and email people stuff. Whatever you decide to do, you will probably want to convert the JSON into some form of readable text output. In order to facilitate this each event type (createCard, commentCard etc.) has an associated Django template. The CallbackEvent model has a render method that passes the entire JSON payload into the relevant template as the template context, so that you can extract the data.

Below is an example of the default commendCard.html template.

<b>{{action.memberCreator.fullName}}</b> commented
on the card "<b>{{}}</b>"
on the board "<b>{{}}</b>":

The default templates are designed to show what is possible - and it’s recommended that you override these in your application. You can do this using Django template overriding - add your template to your application in the same locaion (/templates/trello_webhooks/<event_type>.html) and declare your app above the trello_webhooks app in the INSTALLED_APPS setting, and your template will be used instead of the default.

The combination of overrideable templates and the callback_received signal mean that you should be able to integrate Trello fully into your app.

NB One word of caution

I have made no attempt to ensure that all events are covered - that’s not really the point. This app will store and forward any event that it receives. In order to make it a little easier to manage unexpected events there is a property of the CallbackEvent that is displayed in the admin site list view - Has Template. If this is True, then this is an event for which we have a default template. If it’s False, then this is a new one on us - and you are encouraged to play around with adding a new template. Do please feed all new default templates back to the project.


There are three mandatory environment settings (following the 12-factor app principle):




The first two are the core Trello developer API keys - available from here:

The CALLBACK_DOMAIN is included as you need to give a fully-qualified domain to the Trello API, and it’s not always possible to infer what that might be - for instance when developing locally, you will need a tunnel from your machine out onto the web using something like ngrok.

When managing hooks via the Trello API a third key is required, and this is user specific - the admin site has a link next to the auth_token field on the form for creating a new Webhook. This uses the Trello API client.js to perform the Oauth dance - and supplies the user token. All webhooks are registered against a user token. That’s how it works. (NB you can pass any user tokens you have lying around to the sync_webhooks command and it will check Trello for any existing webhooks registered with those tokens.)


You can run the tests yourself in the normal manner:

$ python test

However, if you have tox installed (and I’d really recommend you do), then you can simply run $ tox, and this will also include coverage.

Coverage isn’t 100% (when is it), but if you do contribute please do include tests for any changes that you make.

The tests themselves use mock objects to replicate the two Trello API calls (list_hooks and create_hook), so no internet access is required. (The project relies on py-trello, and that has coverage for the API calls.)


The app is available on PyPI as django-trello-webhooks, so install with pip:

$ pip install django-trello-webhooks

Further Developments

  • Write some tests

  • Better integration with the Trello API

  • Handle user auth token expiry properly

  • Integration with Heroku’s “Deploy to Heroku” button


Usual rules apply - fork, send pull request. Please try and adhere to the existing coding style - it may not be your style, but it’s the project’s style, so PRs will be rejected if they ‘smell bad’. Specifically, given that this is an app that is pushing data over the wire, and therefore hard to debug - lots of logging, and lots of comments. Seriously. Lots.


MIT (see LICENCE file)


The core Trello API integration is done using py-trello from Richard Kolkovich (@sarumont), so thanks to him for that. He naturally relies on requests from Kenneth Reitz, as well as request-oauthlib, so thanks to anyone involved with either of those.


The webhook API works on the concept of a Trello model id. This refers to the object being watched - and could be a Board, a List, a Card etc. Getting these ids is a bit of a pain, to put it mildly, so I would strongly recommend using the excellent Trello Explorer app from Harald Wartig (@hwartig).

I would also recommend the use of ngrok to expose your local Django dev server during development.

As for development itself - use virtualenv, install dependencies from requirements.txt and set up environment variables. If that doesn’t mean anything to you - I’m afraid you have a lot to learn.

Deploying to Heroku

TODO: write proper docs

This repo contains a test app can be deployed directly to Heroku using their Deploy button. This app will pipe Trello updates directly to a Hipchat room. You will need the following information in order to set up and configure the app:

TRELLO_API_KEY, TRELLO_API_SECRET, which you can get from here - HIPCHAT_API_TOKEN, HIPCHAT_ROOM_ID, which you can get from

In addition, you will need to set the CALLBACK_DOMAIN environment setting once the app has been deployed. This should be set to the <app_name> domain, that is available once Heroku has deployed it.

The recommended hacking method (IMO) is to set up the Heroku app, and use that as your main git remote - pull it down locally, change the relevant templates, push back to Heroku. If you’re actually adding functionality, then please follow the contributing instructions above.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

django-trello-webhooks-0.3.tar.gz (21.9 kB view hashes)

Uploaded source

Built Distributions

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page