Describe the problem/error/question
I am creating a RAG agent using the Q&A Chain Node. I have attached an Ollama Chat Agent with a locally-hosted Deepseek R1. I have used the Vector Store Retriever node, which is connected to a local Qdrant Vector Store. I am using an Embeddings Ollama model for the Qdrant Store. I have a code node to send data to the Q&A Chain Node. I see it is correctly mapped and input to the Q&A Chain.
The output of the Q&A Chain is then connected to another code node (called Data Formatting for Redis storage).
I had it working last week and then worked on some other parts of the workflow. In doing some testing yesterday afternoon, I started to have problems with this agent. Specifically, no matter what I do, I keep getting the error: “Invalid Input Type” for the Embeddings Model.
More puzzling, when I look at the Input mapping in the Chat Agent, Vector Store Retriever, Qdrant Vector Store, and Embeddings Model nodes, all of them detect the Data Formatting for Redis storage code node (the node that FOLLOWS the Q&A Chain] as their INPUT node.
I’ve looked everywhere, but haven’t found any posts of similar problems. Cursor, OpenAI, and Perplexity have also been of no help.
Any ideas?
What is the error message (if any)?
Error in sub-node ‘Embeddings Ollama1‘
invalid input type
Please share your workflow
Information on your n8n setup
- n8n version:1.85.4
- Database:Qdrant
- n8n EXECUTIONS_PROCESS setting (default: own, main):
- Running n8n via:Docker
- Operating system:Mac OS Sequoia 15.4