Configuration and Settings

The Settings interface in BalkanID Copilot allows you to configure AI models and system preferences to optimize your experience with the platform.

Overview

Settings provide comprehensive control over your AI experience through multiple key areas. AI Model Management allows you to configure and manage language model instances, giving you flexibility in choosing the right AI capabilities for different tasks. Model Provider support enables you to add different AI providers, ensuring you can leverage the best available services for your specific needs.

Default Configuration capabilities let you set system-wide preferences that optimize your workflow, while Performance Optimization features help you tune model settings specifically for your use cases. This comprehensive approach ensures that BalkanID Copilot adapts to your organization's requirements and preferences.

Accessing Settings

Accessing the settings interface is straightforward and designed for easy navigation. Click the Settings button, marked with a ⚙️ gear icon in the top navigation bar, to open the settings dialog. The interface opens directly to the Models configuration panel, where you'll perform all model management tasks. This centralized approach ensures that all your AI configuration needs are accessible from a single, intuitive interface.

Other Settings allow a user to view the tenant's current data, and perform syncs on the data.

AI Model Management

Understanding Model Instances

Model Instances represent configured AI models that you can use throughout the platform, each consisting of several key components that define their functionality and accessibility. The Provider represents the AI service provider such as OpenAI, Azure, or Anthropic that hosts the model. The specific Model refers to the particular AI model from that provider, such as GPT-4 or Claude. Configuration encompasses the custom settings and API credentials needed to access the model, while the Nickname provides a human-friendly name for easy identification in selection menus.

Model Instance Properties

Each model instance includes essential basic properties that define its behavior and availability within the platform. The Nickname serves as the display name you see in model selectors throughout the interface, making it easy to identify the right model for your needs. The Provider indicates the AI service provider, such as OpenAI or Azure OpenAI, that hosts the model. The Model ID specifies the exact model version, like gpt-4 or claude-3-sonnet, ensuring you know precisely which AI capabilities you're accessing.

The Enabled status determines whether the model is available for use in the platform, while the Default designation identifies the primary model used when no specific model is explicitly selected. Understanding these properties helps you make informed decisions about model configuration and usage.

Provider Types fall into several categories that help you understand the source and management of different models. BalkanID Providers represent pre-configured models managed directly by BalkanID, offering reliable, tested configurations. Custom Providers are models you configure using your own credentials and settings, providing flexibility for organization-specific requirements. Official Models represent recommended configurations from BalkanID that follow best practices for optimal performance.

Viewing Existing Models

The model list display provides comprehensive information about each configured model to help you understand your available options. Each model shows its nickname and model name for clear identification, along with provider information indicating which service provides the model. Status badges help you quickly identify important characteristics, including BalkanID badges for BalkanID-managed models and Default badges for the currently selected default model.

Adding New Models

Provider Selection

The system supports a comprehensive range of AI providers to meet diverse organizational needs and preferences. OpenAI provides access to GPT models including GPT-4 and GPT-3.5, offering industry-leading conversational AI capabilities. Azure OpenAI delivers Microsoft-hosted OpenAI models with enterprise-grade security and compliance features. Anthropic offers Claude models known for their thoughtful and nuanced responses. Google provides access to Gemini and other Google AI models with unique capabilities and performance characteristics. Additionally, the system supports custom providers for organization-specific or specialized AI services.

If your usecase requires any other AI providers or models to be supported, please contact [email protected], with your specifications, and we will do the needful to support your provider or model.

The provider configuration process follows a straightforward workflow designed to ensure successful setup. Begin by clicking the Add Model button to start the configuration process. Select your desired Model Provider from the dropdown menu, which will populate available options based on the provider's offerings. Choose the specific Model from the available options that best matches your intended use case. Set a descriptive Nickname that will help you and your team easily identify this model in selection menus. Finally, configure the Provider Settings including API keys, endpoints, and other provider-specific parameters.

Managing Existing Models

Editing Models

The model editing process provides flexibility for maintaining and updating your AI configurations as needs change. To edit a custom model, click the Edit button on any model you have permission to modify, then proceed to modify configuration settings as needed. You can update credentials or settings to reflect changes in your provider accounts or organizational requirements, and click Done to save your changes when complete.

Setting Default Models

Default model management ensures consistent AI behavior across the platform when users don't specify a particular model. The system maintains that only one model can be designated as the default at any time, ensuring predictable behavior. The default model is automatically used when no specific model is explicitly selected in any platform feature. To change the default model, simply click Make Default on your preferred model configuration.

Enabling and Disabling Models

Model state management provides fine-grained control over AI availability without losing configuration investments. When a model is Enabled, it appears in selection lists throughout the platform and can be used for AI-powered features. When Disabled, the model is hidden from users but its configuration is preserved, allowing you to quickly re-enable it when needed. The toggle switch provides instant state changes without requiring complex reconfiguration.

Deleting Models

The model deletion process provides permanent removal of model configurations when they're no longer needed. To delete a model, click Edit on the model you want to remove, then click the Delete button in the edit interface, and confirm deletion in the confirmation dialog that appears.

Using a Hosted Model

The BalkanID Copilot supports using a hosted model that is provided by the client. If you want to configure the Copilot to use a hosted model, please contact [email protected] with your specifications, and we will do the needful to support your model.