Everything looks to be ok, Have you tried using the free version of Zep cloud to check if it is a config issue with self hosted or if it is a general issue?
Can you also take a quick look in your Zep logs to see if there are any errors?
Hey Jon,
Thanks for the reply. I tried Zep cloud and am still getting the same error.
By Zep logs, do you mean in n8n or on the Zep backend? Zep logs in n8n look good. It’s unclear how to check for logs in the Zep backend, cloud or self hosted.
I’ve also isolated the Zep node, and tried using a chat trigger instead of my actual workflow, and I am still getting the same error, when i type something simple like “hello” into the chat trigger. I also tried using OpenAI embedding instead of HuggingFace and still getting the same error.
Furthermore, I just tried it with SupaBase, and am still getting the same error.
However, it does work with Pinecone.
We really want to be using Zep though, not Pinecone. Can you help us troubleshoot what the issue is? I suspect there is something wrong in the n8n backend, but I am unsure how to investigate or where to look.
I have this working by using the workflow below, I don’t have a test huggingface instance to use so I used OpenAI but it seems to work and I get the below.
Which error are you seeing? For me the workflow I shared worked with no issues.
I will try with your Zep config as well to see what happens, can you share the url you used for Zep or maybe share if you are including a port or if it has a trailing slash?