setup google-gemini-cli (free-tier, no api) ... help please

[empty nimbus]

hi ... can anyone help with setting up access to google-gemini-cli (free-tier, no api) models?

i have this setup in config.yaml

model:
  default: "gemini-3-flash-preview"
  provider: "google-gemini-cli"
  base_url: "https://cloudcode-pa.googleapis.com/v1internal:generateContent"

...

auxiliary (all provider: auto)

here is error message

⚠️  API call failed (attempt 1/3): CodeAssistError [HTTP 429]
   🔌 Provider: google-gemini-cli  Model: gemini-3-flash-preview
   🌐 Endpoint: cloudcode-pa://google
   📝 Error: HTTP 429: Gemini quota exhausted (You have exhausted your capacity on this model. Your quota will reset after 

i notice that Endpoint has changed from
base_url: "https://cloudcode-pa.googleapis.com/v1internal:generateContent" -> 🌐 Endpoint: cloudcode-pa://google
although this change doesn't show in config.yaml

help please

[scenic sparrow]

cloudcode-pa://google is expected for the google-gemini-cli provider. Hermes uses that as an internal marker for the Google Code Assist/OAuth route; the adapter then sends the actual HTTP traffic to Google's cloudcode-pa.googleapis.com endpoints.

The active failure is the 429:

Gemini quota exhausted

That is coming back from the Google Code Assist backend, not from the display name of the endpoint.

Run /gquota while google-gemini-cli is the active provider to check the remaining daily quota/reset state. If that quota is exhausted, the free OAuth path will not work again until it resets. The alternatives are to wait for reset, switch to the AI Studio API-key provider (gemini), or use another provider.