Feasibility of sending video files into n8n’s AI Agent node (using Google Gemini 2.5 Pro) and possible workarounds or roadmap plans

Questions
1

Hi everyone,

I’d like to explore whether and how it’s possible to feed video files directly into the AI Agent node in n8n, using Google’s Gemini 2.5 Pro model for analysis. From my testing so far:

  • AI Agent node today only supports automatic passthrough of binary images, not videos.
  • Gemini 2.5 Pro via the Google Vertex AI API does accept video inputs, but n8n’s built-in Agent node has no checkbox or toggle for video binaries.
  • I’m curious if there’s any upcoming support for binary video in the AI Agent node, or if the n8n team plans to add that soon.

My questions for the community:

  1. Has anyone successfully sent video files through n8n into the AI Agent node?
  2. Are there recommended workarounds ?
  3. Does the n8n roadmap include first-class video support (analogous to image passthrough) for Agents?
  4. Any pointers to open GitHub issues or feature requests I should upvote?

Cheer,