Model Providers
The Model Providers page (/admin/model-providers) lets workspace administrators configure and manage AI model providers. Only admin users can access this page.
Supported providers
BasePeak.AI supports a variety of model providers, including:
- OpenAI
- Anthropic
- xAI
- Ollama
- Voyage AI
- Groq
- vLLM
- DeepSeek
The UI shows whether each provider has been configured. Configured providers display the number of available models and offer options to manage models or remove the provider.
Configuring a provider
To add and configure a provider:
- Click Add Provider to open the provider picker
- Select the provider type and give the instance a display name
- Click Configure on the new instance
- Enter the required credentials (API key, endpoint URL, etc.)
- Save — BasePeak.AI validates the credentials immediately
You can configure multiple providers, letting you pick the right provider and model for each use case. Read-only providers bootstrapped by the BasePeak platform (such as the managed default provider) cannot be reconfigured through the UI.
Viewing and managing models
Once a provider is configured, open its Models view to manage the models it offers. For each model you can:
- Activate or deactivate the model (controls availability to users)
- Set the usage type to tell the platform how the model should be used:
| Usage type | Description | Used in |
|---|---|---|
| Language Model | Text generation and tool calls | Agents and tasks; selectable as an agent's primary model |
| Text Embedding | Converts text into numerical vectors | Knowledge tool (RAG) |
| Image Generation | Creates images from text descriptions | Image generation tools |
| Vision | Analyses and processes visual content | Image vision tool |
| Other | Default when no specific usage is selected | Available for all purposes |
Use Refresh models from the provider action menu to pull the latest model list from the provider.
Setting default models
Click Set default models (top-right of the Model Providers page) to choose platform defaults for:
- Language Model (Chat) — primary conversational model
- Language Model (Chat - Fast) — optimised for quick responses
- Text Embedding (Knowledge) — used for knowledge base operations
- Image Generation — for creating images
- Vision — for image analysis
These defaults apply when a user or agent has not selected a specific model. After choosing the defaults, click Save Changes.