# Supporting a new connector without webhooks

**URL:** <https://forum.rasa.com/t/supporting-a-new-connector-without-webhooks/9125>\
**Category:** Rasa Open Source\
**Created:** [April 29, 2019, 10:59am UTC](https://forum.rasa.com/t/supporting-a-new-connector-without-webhooks/9125 "2019-04-29T10:59:48Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Bij](https://avatars.discourse-cdn.com/v4/letter/b/7cd45c/32.png) [@Bij](https://forum.rasa.com/u/Bij)\
**Post date:** [April 29, 2019, 10:59am UTC](https://forum.rasa.com/t/supporting-a-new-connector-without-webhooks/9125/1 "2019-04-29T10:59:48Z")

</div>

Hi

I would like to implement a custom rasa channel - but a webhook is a total no-no (unsafe for information security reasons). I’d prefer to use the Slack Real Time Messaging API - [Real Time Messaging API | Slack](https://api.slack.com/rtm) rather than webhooks.

I see no examples, and I’m unsure a generic style like this is supported without some aggressive code changes to the platform.

If you can suggest anything - LMK! 😀

---

<div class="post-metadata">

**Author:** ![tmbo](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/tmbo/32/19_2.png) [@tmbo](https://forum.rasa.com/u/tmbo)\
**Post date:** [May 6, 2019, 7:33pm UTC](https://forum.rasa.com/t/supporting-a-new-connector-without-webhooks/9125/2 "2019-05-06T19:33:19Z")

</div>

Hey @Bij and welcome to the Rasa community 🙂

The current architecture is def build around a push model and not a pull one (which would be the messaging API, which uses a websocket to retrieve messages). Why do you think a pull would be more secure? I mean you should use HTTPS in either direction otherwise you are right.

---

<div class="post-metadata">

**Author:** ![Bij](https://avatars.discourse-cdn.com/v4/letter/b/7cd45c/32.png) [@Bij](https://forum.rasa.com/u/Bij)\
**Post date:** [May 6, 2019, 10:09pm UTC](https://forum.rasa.com/t/supporting-a-new-connector-without-webhooks/9125/3 "2019-05-06T22:09:00Z")

</div>

Unfortunately in enterprise environments - you don’t always control the process. I do think that most enterprises don’t expose endpoints to the internet- so I don’t see how this will impact adoption of the container system proposed

I do think that you are going to hit this use-case for enterprise customers and if I was in your shoes I’d certainly be thinking about it. At minimum an asyncio pump that people can use to perform background tasks. I had to hack this by creating a subprocess polling /pump
