Users are able to upload files during a conversation with an agent. This feature is useful when users need to share files with agents or with human operators.

Files uploaded during a conversation are not accessible to other conversations. To extend the Agent’s knowledge for all conversations, use a Datastore instead.

example 1

Agents are able to process the following file types:

  • text/csv
  • text/plain
  • text/markdown
  • application/pdf
  • application/json
  • application/vnd.openxmlformats-officedocument.presentationml.presentation
  • application/vnd.openxmlformats-officedocument.wordprocessingml.document
  • application/vnd.openxmlformats-officedocument.spreadsheetml.sheet
  • image/png (Vision compatible model like GPT-4-Turbo or Claude 3 required)
  • image/jpeg (Vision compatible model like GPT-4-Turbo or Claude 3 required)
  • image/gif (Vision compatible model like GPT-4-Turbo or Claude 3 required)
  • image/webp (Vision compatible model like GPT-4-Turbo or Claude 3 required)
Unsupported file types will be ignored by the AI

To reuse an uploaded file use the conversation files dropdown example 2

Was this page helpful?

Fine tuningAnswer sources