Hi all,
I’ve been dealing with a persistent platform failure for over a month now, and I know I’m not the only one. ChatGPT-generated file download links—specifically anything using the sandbox:/mnt/data/ path—are completely non-functional across .docx, .pptx, .zip, and other formats.
Here’s the situation:
I’ve tested across browsers, devices, and network environments
I’ve submitted multiple support tickets
I’ve directly emailed OpenAI support and CC’d leadership (including Brad Lightcap, Mira Murati, and Logan Kilpatrick)
I’ve received no technical explanation, no ETA, and no fix
The issue has remained completely unresolved for weeks
This isn’t just a minor inconvenience—it breaks a core feature for professional users who rely on file generation and download. The most frustrating part? No one from OpenAI has provided a real update. Support continues sending placeholder responses like “we’re investigating.”
I’ve documented every failure with timestamps, and I’ve spoken with other users experiencing the same. This is not isolated.
If you’re also affected, please comment to confirm. We need OpenAI to:
-
Acknowledge the issue publicly
-
Provide a root-cause explanation
-
Offer a workaround or ETA
-
Commit to better reliability and transparency moving forward
Thanks for your time—and to those at OpenAI who are genuinely trying to help, we’re counting on you to surface this where it matters.