local-ai-packaged Chat workflow does not use Qdrant data and python error

[dan-burt]

I followed the guide for the project using the Apple Silicon M1 pathway. I have Ollama manually installed, with the 2 models coded into the default config, as well as a few others.

I create the 3 credentials as instructed, but I change the workflow to watch a local (to the n8n container) folder. I can see this data in the Qdrant UI.

When I chat with the model, 2x pictures of the workflow to show that it is reading in the correct data.

1st image shows it summarising a list of events in the text file. There is a returned error as well.

< python_tag >tool call: get_christmas_plans()

2nd image shows the result of the Chat workflow's Ollama Model node, in case it showed any more detail(?)

[mp3pintyo]

Unfortunately, Ai Agent, Qdrant, and Ollama do not work well together.
https://community.n8n.io/t/ai-agent-vector-store-tool-generates-strange-responses-and-seems-to-miss-the-tool-information/54952/7