I’m running n8n in the cloud version, and this example flow ( RETIEVER ) worked perfectly for me, but I realized that it’s not saving chat history ( input output ).
I tried a lot to add it and couldn’t, so I asked for your help.
**n8n version:**1.72.1
Database (default: SQLite):
n8n EXECUTIONS_PROCESS setting (default: own, main): main
good morning
I think its because you use a question and answer chain and this has no memory. As far as I understood the advantage is that its not making up stuff and only replies what is in your documents in the vector store.
If you use an AI Agent instead for example an tool agent you can connect a memory for example via postgres chat memory and save the chat in supabase and use your supabase as a vector store tool to have access to your documents.
I´m also quite new to that but I think it could work like that