Connect an OpenAi Assistant

Hello all, I am using the OpenAi plugin in my project and I had successfully manage to get a response and use it in my app.
Now, I created an assistant in the OpenAI playground, but I can’t see how to use the assistant in Weweb. Is that possible or can we only access the chat function ?

Hi Denis! Currently, you cannot connect to the new GPT assistants using our OAI plugin. Those assistants are still in Beta mode; they still lack many features (streaming, for example) and will go through a relatively rapid development before they are seen as “production-ready” by OAI.

Once they do lose the Beta sticker, we will accommodate that in our plugin.

In the meantime, you can connect to the assistants using their API, which is well-documented (https://platform.openai.com/docs/assistants/overview) and relatively easy to use once you understand how the assistants work.

My personal view is that the GPT assistants are amazing, and they will enable incredible use cases. I have a playground in Xano where I experiment with them, and it is super easy.

1 Like

Thanks Slavo for this detailed answer. As I have no particular deadline, I will wait until these assistants get in production-ready status at OpenAI and implemented on your end. I agree, these assistants sounds very promising.

Hi Denis,

Zapier has a setup which connects to Assistant and a chat-interface which I think can be integrated in the web. Some inspo/intro: https://youtu.be/IPdL5QOuPMo?si=vhU3ZJS8r8BOGIU9

I plan to integrate it later this month :slight_smile:

Hey Slavo do you have this in a Xano snippet I would love to play around with it too hahaha

1 Like

Hi! Sorry, no, I don’t, it is just a few functions and calls. I really didn’t have time to do anything serious with it; I just wanted to check how to go about it.

Who did more then just playing around with the assistant - looking to hire someone that can help us set it up for one of our use cases. Thank you!

Hi Raelyn,

Did you manage to play around / find a Xano snippet?

Thanks

Nope! But it should be fairly easy to call any open AI end point with Xano and display the results in Weweb :blush: