#Uploading PDF documents to GPTs fails

1 messages · Page 1 of 1 (latest)

edgy grail
#

Bug Report:

Steps to reproduce:

  1. Create a new GPT or edit a already created GPT
  2. Switch to the Configure tab in edit mode.
  3. In the Knowledge section of the screen: Click the Upload files button and select a PDF document (e.g. one created by saving as PDF from Google Docs)

Expected result:
Document shows up in the list of documents miniatures and some success message is displayed somewhere that the upload was successful. Clicking on the miniature opens a preview of the document.

Actual result:
An error message is shown in two places after the upload seems to be complete - in contrast to this error a miniature for the PDF is added. The orange error messages disappears (too quickly) and the one on the top right is persistently shown.
Clicking on the miniature of the new PDF does not open a preview.
When leaving edit mode and returning into edit mode the miniature has vanished.

Additional information

Browser: Safari 17.0 as well as Brave Browser v1.60.114 (Nov 8, 2023) (Chromium based)
OS: macOS Sonoma 14.0

Link to the GPT can be provided to the Dev team if required.

Problem ist not restricted to PDFs, any file type fails to upload. Itried PNG, JPG as well.

Problem also impacts the document upload which is possible in edit mode in the Create tab - same error messages appear on the screen and eventhough the model references the document contents after uploading, is seems that it is not saved / stored, as when leaving and entering the edit mode, the document is not visible anymore in the Knowledge section of the Configure tab.