Talk To Your Data (Preview) - Configuration

Once the backend is running and the Talk To Your Data interface is accessible, follow these steps:

Use your administrator credentials to log into the Talk To Your Data Web User Interface.

2. Navigate to Settings

Go to Settings → Admin → Models:

Settings page showing AI Models and Providers configuration, with gpt-4.1 selected as the default model under Azure OpenAI and an associated API provider listed below. — AI Models & Providers settings

3. Add API Providers

Before adding any models, go to API Providers and add your required AI providers (for example, AzureOpenAI). This typically includes setting a valid endpoint URL and API key:

Add New Provider form for Azure OpenAI, showing required fields for provider name, Azure endpoint URL, API key, and API version. — Add API Provider dialog

4. Add Models

Click Add, select the provider and specify required model, save the configuration:

Add New Model form showing fields for model display name, API model name, provider selection, and model visibility and default settings. — Add New Model dialog

5. Add Model for Embeddings

Click Add Embeddings Model, configure by providing its URL and API key under the same API Provider setup. This model is used to process and interpret user questions.

Add Embedding Model form showing fields for display name, provider API identifier, provider selection, embedding dimension, and advanced token limit settings. — Add Embedding Model dialog

6. Ready to Use

Once all models are configured, you can begin interacting with your data using natural language.

Make sure that appropriate access permissions are granted to the user for any data sources, views, or tables that should be available through the interface. Without proper permissions, the AI will not be able to query or return results from the intended data.

1. Log in

2. Navigate to Settings

3. Add API Providers

4. Add Models

5. Add Model for Embeddings

6. Ready to Use