Voice replication in Google AI Studio fails after recording the voice sample and consent statement and clicking Replicate voice.
Model: gemini-3.8-flash-tts
The failed MakerSuiteService/CreateVoice request returns HTTP 404 with this complete response:
[5,"Requested entity was not found."]
I verified that:
- The project is on paid Tier 1 with prepaid credit available.
- Gemini API is enabled.
- The API key is available and restricted to Gemini API.
- My account has the project Owner role.
The error persists with a newly created project and API key.
Does Gemini 3.8 voice replication in AI Studio require additional approval or allowlisting? If not, is this a known issue, and how can I identify which resource wasn’t found?