How Rocket.Chat + ChatGPT Integrations Work with Moqui

I have an idea about :How Rocket.Chat + ChatGPT Integrations Work with Moqui.
I think these two tools like to giving Moqui system a pare of wings to fly.

Step 1: Authenticate Rocket.Chat and ChatGPT.
Step 2: Pick one of the apps as a trigger, which will kick off your automation.
Step 3: Choose a resulting action from the other app.
Step 4: Select the data you want to send from one app to the other.
That’s it! More time to work on other things.


A possible integration workflow might look like this:

  1. A user interacts with Rocket.Chat, and the message is sent to a designated channel or bot within Rocket.Chat.
  2. The Rocket.Chat bot processes the message and decides that it needs assistance from ChatGPT.
  3. The Rocket.Chat bot sends the user’s message to ChatGPT using the OpenAI API and awaits the response.
  4. ChatGPT processes the request, generates a response, and sends it back to the Rocket.Chat bot via the API.
  5. The Rocket.Chat bot receives the response from ChatGPT and posts it in the chat channel for the user to see.
  6. If needed, the Rocket.Chat bot can also take action within Moqui based on the conversation. For example, it might trigger a workflow or update data in Moqui based on the user’s request.

Interesting. This would be cool. Have you implemented it yet?

Still under investigation and evaluation, hubot may need to be introduced as a media

1 Like

After a period of research, the original direction needs the following adjustments: introducing Rasa as a bot to interact with Moqui and integrating the currently popular LangChain project. There is a sample project on GitHub called RasaGPT, which also needs further investigation. The original Hubot bot is outdated. Rocket.Chat is a private domain chat tool that can be considered for integration. Alternatively, integration with chat tools like Telegram or WeChat can also be considered.

So, the current approach should be Moqui + Rasa + LangChain + Rocket.Chat.

1 Like

It’s great to hear your progress.

Do you have a demo or code? I’d like to see what you’ve done if possible

Please be patient for a while longer.
I will gradually upload the code to GitHub.
At this stage, I have only conducted extensive research and verification of the development tools.
In the future, I plan to release code examples related to integrating Rasa with Moqui.

No problem. I’m just interested and intrigued. Keep up the good work!

I have spent a considerable amount of time researching the AI landscape to find the most suitable AI model to integrate with Moqui. The focus here is on using a question-and-answer interaction mode. Ultimately, the choice will need to be made between two seemingly similar but significantly different frameworks: Rasa and LangChain.

At this point, it is difficult to say which one is better:

  • Rasa has a relatively mature community environment and a good track record of commercialization. However, it seems that it is not so easy to use in terms of interaction capabilities and debugging complex problems.
  • LangChain is currently very popular and cannot be ignored. However, its commercialization path still needs to be observed for a longer period of time, perhaps it will be clearer after a year.

I have also investigated open source projects similar to RasaGPT, which has been discontinued for 9 months now. This confuses me a bit, because in terms of the Way of thinking, combining the two seems to be a good solution that can be implemented relatively quickly. I don’t know what happened in the middle that led to the current long-term suspension of updates.

Originally, using a Rasa case, I made a Moqui integrated hello world. Users can create user accounts and query user information through chat. However, I encountered many confusions in the process of debugging Rasa, which made me doubt the usability of Rasa itself.

I have noticed that the Rasa community seems slow to adapt to the rapid developments in the current AI field. In light of this, I am considering organizing the previous work integrated with Rasa to provide a preliminary entry point implementation. After that, I plan to shift my focus to closely tracking the development of Langchain. I will reconsider the technical solution selection for the idea of using Rasa to cover the entire Moqui interaction comprehensively. In the next phase, more attention will be dedicated to the integration work related to Langchain. The work on the Rasa project will be shared as code on GitHub. I will provide further updates once the organization is complete.