Mike Nijs
Hey team,
What we're looking for:
  1. Users should be able to paste images directly into the chat input (clipboard paste via Ctrl+V / Cmd+V)
  2. Users should be able to upload files as attachments (drag & drop + file picker button)
  3. An allowlist/blocklist for file types at the chatbot settings level, so we can control per bot which extensions are accepted (e.g. allow .pdf, .jpg, .png, .docx / block .exe, .bat, etc.) -> important for us, is that we can use CAD files (.dxf / .stp / .step).
  4. The uploaded file must be accessible in the (new) Agent node so we can process it downstream (to n8n).
One thing that's important from an integration perspective: the attachment should be part of the same webhook payload as the message (not a separate event or callback). We catch and process everything in n8n (webhook → JSON parsing → downstream nodes for handling, analysis, storage, etc...), so having file and message context arrive together in one execution makes all the difference.
Whether that's a signed URL, base64, or something else is up to you, we can work with whatever lands in the JSON.
Happy to elaborate if helpful.
Cheers,
Mike