I want to use my company’s Confluence wiki as a RAG for the AI Agent.
This is my first attempt at doing this.
So, I’ve created a workflow to search the Confluence by REST API, after I want to use the response in AI Agent, like the workflow below.
This is the best approach?
How I use the markdown as information for the Agent (database or tools) ?
I would store the information in a Vector Store, if it’s too big.
For setting an awesome RAG Agent and Vector Store using Supabase free plan, I recommend these series of videos:
Your question will be answered by analyzing the Agent prompt and the logic behing the database, when storing the documents.
His workflows are available to download for free on GitHub. It’s the best RAG Ai solution I’ve seen with n8n so far.
Alternatively, you could create an Assistant in OpenAI API platform and add the text in the knowledge base of the assistant.
You would then simply use the “Message Assistant” node:
It’s a good start, if you’ve never built agents before. Very easy to setup and start using.
.
If my reply answers your question, please remember to mark it as a solution.
Hi. You would use an agent flow like this. Use your original flow code as a tool for your ai agent to return back the markdown. Instead of a form, use a chat to get the information.
Hi, thanks for answering me.
Your solution works, but at the end of the wokflow called as a tool the following error occurred:
No path back to referenced node
There is no connection back to the node **'Get Availability'**, but it's used in an expression here.
Please wire up the node (there can be other nodes in between).
Make sure you edit this portion of the workflow. Once it doesnt have any errors, copy this portion of the workflow, and paste it inside the Get Availablity tool.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.