Workflow for code documentation support

Questions
1

Describe the problem/error/question

I am new to n8n and would like to create a workflow to answer support questions for a library. My idea is to be able to get the question from Slack and have .md files for each component with its documentation from my GitHub repo, use ChatGPT to process the data, and send a response. I don’t quite understand how to set up the workflow because I need to be able to distinguish whether the question is about the Android or iOS platform so I can decide which GitHub repository to search for the documentation. I have something like this in mind:

Slack Question → Platform Detection → Documentation Search → AI Response → Slack Reply

 ↓                         ↓                                      ↓                                              ↓                     ↓

Trigger Filter Node File/GitHub ChatGPT Slack Node

Is there a template that can guide me, or any recommendations you can give me?

2

Hey @Edgar_Navarro hope all is well. Welcome.

What you can do is you can build a RAG system, with platform tag in metadata and then in the conversational AI Agent, interrogate the user for the system they are asking about (when not possible to deduce from context), then use vector datastore to pull related info and compile an answer).

Use the following video if you want to learn about RAGs:

3

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.