- RubiAI Model Configuration allows administrators to connect Rubiscape with Large Language Model (LLM) providers such as Gemini.
- Once configured, RubiAI features become available across Rubisight and Rubistudio for Designer Assistant, widget insights, commentary generation, and Prompt Studio.
- This is a tenant-specific configuration and must be set once per tenant.
Accessing the RubiAI Model Configuration:
- This option is available to Administrators under:
Navigation: Left Menu → Rubi AI Model Config
- The configuration screen displays provider options, API key fields, model selection, and tuning parameters.
1. Select an LLM Provider
- The “Select a Provider” dropdown lists all available Large Language Model providers.
- Currently supported provider: Gemini.
- Select the provider you wish to use for your tenant.
2. Add API Key
- Enter the API key associated with your LLM provider account.
- This key enables the system to authenticate and retrieve model information.
- Ensure that the API key is valid and active.
- After entering the API key, click “Get Models.”
- The system will fetch and display the list of available models provided by the selected LLM vendor.
- Models may include variants with different capabilities such as flash, pro, reasoning, or high-context versions.
4. Select a Model
- Choose a model from the dropdown as per your requirement.
- Model selection determines:
- Response performance
- Output detail level
- Token limitations
5. Set Model Parameters
- Configure tuning parameters for the selected model:
Temperature:
- Controls creativity and randomness of responses.
- Lower values produce deterministic outputs.
Max Output Tokens:
- Determines maximum token count for generated responses.
Top P:
- Controls sampling diversity using nucleus sampling.
- Adjust sliders as needed based on your use case.
6. Save the Configuration
- Click the “Save” button to apply the settings.
- These settings become active immediately for the tenant.
Behavior After Configuration:
- Once configured successfully, RubiAI features across Rubiscape become enabled, including:
Rubisight:
- Designer Assistant
- Widget narration
- Dashboard commentary
- Widget insights
Rubistudio:
- Prompt Studio
Notes for Administrators:
- This configuration is tenant-specific; each tenant must configure RubiAI separately.
- API keys are provider-specific and should be kept confidential.
- Model parameters can be updated anytime if your use case changes.
- Ensure you have active billing/subscription where required by the LLM provider.
- If the API key becomes invalid or expires, AI features will stop functioning until updated.
Summary:
- Configuring the RubiAI model enables seamless use of AI functions across Rubiscape’s analytics tools.
- Administrators must select a provider, add an API key, fetch available models, select an appropriate model, tune parameters, and save the configuration to activate RubiAI capabilities for the tenant.