Send binary (not base64) to Gemini Files (queue mode)

Questions
1

I’m having issues uploading files to Gemini Files through its API.
First I thought it was a problem of self-hosted n8n using always base64 format, but I converted the file to binary with an Aggregate node and still wouldn’t work.

This is similar to this issue.
And this is the documentation: Files API  |  Gemini API  |  Google AI for Developers

I read this is solved by setting the env N8N_DEFAULT_BINARY_DATA_MODE to filesystem but in queue mode this is not possible. Jan said (I spent some time trying to retrieve the post) in the future Memory mode will be deprecated, and only filesystem will be permitted, I don’t know how they will handle this in queue mode.

Information on your n8n setup

  • **n8n version:**1.89.2
  • **Database (default: SQLite):**postgres
  • **n8n EXECUTIONS_PROCESS setting (default: own, main):**queue
  • **Running n8n via (Docker, npm, n8n cloud, desktop app):**npm
  • **Operating system:**Ubuntu 24.04
2

Hi,

There was a similar question here:

Not sure if there is a big difference in queue mode

reg,
J.

3

I tried and still failing.

I’m debugging by simply using curl and also get some issues there, so it might be something else. It’s strange because the getURL HTTP Request node works, but the actual upload doesn’t. When it seems it should work, it simply idles loading until I stop it.

I might make some community nodes for Google Gemini, this model is barely supported in n8n.

4

Hello, I have a template that no need for embedding, no need to set up reranking, and no need to connect to a vector database.
It automatically downloads specified files from Google Drive as a knowledge base to answer questions. Files (txt, complex PDFs, images, or any format) can be used directly as the knowledge base.