Setup for Anthropic, Ollama, and OpenAI-compatible providers — endpoint URLs, API keys, model selection
Connect to Claude models via the Anthropic API. Requires an API key from console.anthropic.com. Configure the endpoint URL and select which models to make available to users.
Run AI models entirely on your own hardware with Ollama. No API key required — the Workbench connects to your local Ollama instance. This provides total data privacy: no prompts or data leave your machine.
Connect to any provider with an OpenAI-compatible API: DeepSeek, OpenAI, Azure OpenAI, and many others. Configure the base URL and API key. The Workbench auto-discovers available models from the provider's /models endpoint.
Admins control which models are available to users. Each provider's models can be individually enabled or disabled. Users select from the enabled models when using the Chat Agent, AI-guided recipes, or other AI features.
| Provider | Type | API Key Required | Privacy |
|---|---|---|---|
| Anthropic | Cloud | Yes | Data sent to Anthropic |
| Ollama | Local | No | All data stays local |
| OpenAI-compatible (DeepSeek, etc.) | Cloud | Yes | Data sent to provider |