I’m unable to upload images on OpenAI’s Prism research platform. Is anyone else experiencing the same issue?
Welcome to the forum!
Thanks for repoting.
I have not used OpenAI Prism before but wanted to verify if the problem is reporoducable.
Using OpenAI Prism via an internent browser with URL starting wtih https://prism.openai.com/?u=
in the middle panel at the bottom find
when clicking shows
I do not get the option to upload. However I am also not signed-in to an OpenAI account.
After signing-in
Using Windows 11, clicked on a image in this forum,
then pasted it into the Chat box.
No errors reported.
Then deleted the image from the Chat box.
Copied the image to my Desktop as Coffee machine.png
Then using Upload files & photos
No errors reported.
Can you give step by step instructions so we can try to reproduce the problem.
Also use F12 key to open the developer tools and look for errors reported.
In doing the same using Chrome, these are noted.
/api/project-files/upload-plan:1 Failed to load resource: the server responded with a status of 502 ()
b63323c505d5b182.js:1 Cross-Origin-Opener-Policy policy would block the window.closed call.
(anonymous) @ b63323c505d5b182.js:1
b63323c505d5b182.js:1 Cross-Origin-Opener-Policy policy would block the window.closed call.
(anonymous) @ b63323c505d5b182.js:1
b63323c505d5b182.js:1 Cross-Origin-Opener-Policy policy would block the window.closed call.
(anonymous) @ b63323c505d5b182.js:1
b63323c505d5b182.js:1 Cross-Origin-Opener-Policy policy would block the window.closed call.
(anonymous) @ b63323c505d5b182.js:1
b63323c505d5b182.js:1 Cross-Origin-Opener-Policy policy would block the window.closed call.
(anonymous) @ b63323c505d5b182.js:1
b63323c505d5b182.js:1 Cross-Origin-Opener-Policy policy would block the window.closed call.
(anonymous) @ b63323c505d5b182.js:1
/_next/image?url=https%3A%2F%2Foaisdmntpreastus2.blob.core.windows.net%2Ffiles%2F00000000-dd68-71f6-8f1d-444874f61fec%2Fraw%3Fse%3D2026-05-14T13%253A40%253A44Z%26sp%3Dr%26sv%3D2026-02-06%26sr%3Db%26scid%3D1d1ad5cd-e4c9-4873-9726-6cac73efcea9%26skoid%3D02b7f7b5-29f8-416a-aeb6-99464748559d%26sktid%3Da48cca56-e6da-484e-a814-9c849652bcb3%26skt%3D2026-05-14T13%253A29%253A55Z%26ske%3D2026-05-16T13%253A29%253A55Z%26sks%3Db%26skv%3D2026-02-06%26sig%3Ddg%2Fy8vFEz5BDgaEUFRj787ToWfbEfr6Zha7OBlXk4WY%253D&w=1920&q=75:1 Failed to load resource: the server responded with a status of 400 ()
In the lower left there is a way to contact support@openai.com
I can successfully upload images using the method you suggested. However, when I take a screenshot directly on my computer and paste it via Ctrl+C / Ctrl+V, the upload fails and displays garbled text, exactly as shown in the error screenshot I shared earlier. Could you please help clarify what might be causing this?
First, just to be clear, I am not an OpenAI employee, so I only have access to the same public information and forum details that you do.
That said, one possible cause may be the image format.
In the example where the image was copied with Ctrl+C and then pasted with Ctrl+V, the uploaded image appears to be a JPEG:
That may or may not be relevant, but it is one detail worth checking.
Could you provide the exact steps you used to reproduce your current problem, similar to the steps I listed above?
Without precise reproduction steps, it becomes difficult to help because too many assumptions have to be made. Details such as the browser, operating system, image source, image format, and whether the image was copied, dragged, uploaded, or pasted can all matter.
Thank you for the reminder. I’ll look into it. I suspect it’s probably due to formatting issues when copying and pasting. By the way, could you let me know what image formats Prism supports?
Just a reminder: I am not an OpenAI employee, so I only have access to the same public information you do.
One place to check is the OpenAI Help Center. There is a Prism-related troubleshooting page here:
https://help.openai.com/en/articles/20001050-troubleshooting-and-getting-help-in-prism
@OpenAI_Support Could you answer this directly, and possibly update the Prism FAQ with the supported image formats? Thanks.