I am using the Gemini API with the Files API for video understanding on large files on behalf of users in a unique case.
I built a pipeline that handles several concurrently and removes them once inference is complete.
However it seem like I am hard capped to 20GB per Google Console project Gemini API token.
Ideally I would not have to spawn several projects and maintain several tokens.
It would be much easier if I could increase quota for requests and file storage to have higher concurrent capacity.
Is this possible with API tokens? Can someone at Google increase this? It seems rather arbitrary when you can make multiple projects to accomplish the same thing.
I’d rather avoid vertex with google cloud storage since some users might want to bring their own token to manage their own data.
Hi @Evan_Lesmez,
Welcome to the Google AI Forum! 
Can you share which are you currently using and your project tier?
Hi Krish. Which project I am currently using or which model?
I am using API key(s) generated from https://aistudio.google.com/projects.
Every project I have created is listed as Tier 1 for quota.
I target Gemini-2.5-Pro.
I setup vertexai before with ADC credentials but it was a lot of overhead and like I said some users might want to manage their own data without going into my bucket.
So is this possible or do I need to hack my way around the imposed quotas?
This is incredibly limiting for production usage. I have recently tried to migrate from Vertex API to gemini ai api because vertex is incredibly unreliable with api/tokne rate limits with really random “resource exhausted” messages, but the Gemini API is incredibly non-production ready because the 20GB is very small for working with video / larger files
Hi @Evan_Lesmez ,
-
The 20GB limit is the default starting quota for the Gemini Files API.
-
You must formally request a quota increase to get more storage space.
-
How to ask: Go to the Google Cloud Console, find the “Quotas” page, and search for the “Generative Language API”.
-
What to change: Select the “File Storage” quota and click “Request increase”.
-
Important: Provide a strong reason for the increase in the request form.
-
Tip: Make sure your code deletes the video files immediately after processing them to keep your storage usage low and avoid hitting the limit again.
Welcome to the community, @DeS ,
-
The “resource exhausted” errors you previously faced on Vertex AI are typically due to hitting default rate limits (RPM/RPS), which can be resolved by submitting a quota increase request specifically for the Vertex AI Generative AI API.
-
In the Vertex AI platform and submit a quota increase request for both File Storage and Requests per Minute (RPM) to achieve the necessary production scale.
Thanks for the response @Pooja_Kapse. However I do not see an option to “Request increase”.
I cannot find a way to submit a request to increase.