Hello Google Engineering Team,
I am an NVDA screen reader user, encountering a persistent Permission Denied error across all Gemini models in Google AI Studio. I would like to note that this issue has only started recently, and I have previously been able to use Google AI Studio without any problems.
I have attached a screenshot of the browser console logs, and I hope it contains all the information needed for diagnosis. However, if the screenshot is insufficient, I would be happy to share the full console log text through any alternative channel you can suggest, as the raw logs contain a large number of URLs that exceed the forum’s two link limit for new users.
This issue is cross platform for my account. It occurs on my desktop computer and also reproduces identically when trying to access Google AI Studio via Google Chrome on my mobile phone.
Environment Details:
• Desktop OS: Windows 10 Version 22H2 (Build 19045.6466)
• Desktop Browser: Google Chrome Version 149.0.7827.156 (Official Build) (64-bit) (cohort: Stable)
• Cache/Cookies: I have tested in a fresh browser profile with cleared cache and cookies, but the issue still persists.
Observed Errors by Model Category:
• Speech and Music models: HTTP 400 or 500-level response with error code 403 (Permission Denied).
• Code and Chat models: “Failed to generate content: permission denied. Please try again.”
The error is consistent across all tested model categories, which suggests this is not isolated to a specific model type.
If you need any additional diagnostic information; such as request IDs, HAR files, or the full console log text; please let me know and I will do my best to provide them.
Thank you for your time and assistance.