Account restricted to AQ. tokens after running a RAG chunking loop — requesting AIza key restoration

Gemini API
, ,
1

Hello Google AI Developer Support Team,

I am writing to request a manual review and restoration of my account’s ability to generate standard, production-ready “AIza…” format API keys in Google AI Studio.

Issue Description

All newly generated API keys across my entire account—including inside completely fresh project containers—are being restricted to an “AQ.” prefix token format.

Because of this “AQ.” formatting constraint, my application’s orchestration framework throws an ACCESS_TOKEN_TYPE_UNSUPPORTED (401 Unauthenticated) error when hitting the generativelanguage.googleapis.com endpoints, completely blocking my development environment.

What Triggered the Flag

My account was automatically flagged while testing a local Retrieval-Augmented Generation (RAG) pipeline. I uploaded a 20+ page text document into my orchestration canvas. The framework attempted to vectorize and embed the document chunks rapidly in parallel using the Gemini Embedding model.

This unexpected background burst of concurrent requests tripped your automated abuse/scraping defense system. It was purely an experimental, un-throttled data-chunking pipeline and not an intentional violation of terms.

Current Account Context

  • Google Account Email: [INSERT_YOUR_GMAIL_ADDRESS_HERE]
  • Current Status: All previous keys have been deleted. No keys are exposed publicly.

Could you please review my account and restore my capacity to generate standard “AIza…” API keys so that I can resume building my application?

Thank you for your time and assistance.

2

We are moving away from Traffic keys (Alza) and towards a more secure Authentication Key (AQ). AI Studio will now only generate Authentication Key (AQ) keys. More information on key usage can be found in the Using Gemini API keys documentation.

Please submit any compatibility issues with services / SDKs etc. in this form for our engineering team to review. Thanks!