Gemini is Google's AI model family that can be accessed directly through the Google AI Studio API.

Before you begin

Make sure that your Google Cloud account has a project with the Gemini API enabled.

Configuring Gemini

1. Get your API key from Google AI Studio.

2. Create a Kubernetes Secret that stores the API key.

   kubectl create secret generic kagent-gemini -n kagent --from-literal GOOGLE_API_KEY=<your_api_key>

3. Create a ModelConfig resource using the Gemini provider. The following Gemini models are supported:

   • gemini-pro - Text generation with function calling support
   • gemini-pro-vision - Multimodal model for text and image processing
   apiVersion: kagent.dev/v1alpha2
   kind: ModelConfig
   metadata:
     name: gemini-model-config
     namespace: kagent
   spec:
     apiKeySecret: kagent-gemini
     apiKeySecretKey: GOOGLE_API_KEY
     model: gemini-pro
     provider: Gemini
     gemini: {}