Skip to main content

Model providers

The Model Providers page is where you connect Orbis Studio to the AI model providers your flows use, and manage the API keys for each one. Configuring a provider here makes its language models and embedding models available to every flow in the workspace, so builders can select a model without pasting credentials into individual components.

Open the Model Providers page​

  1. In the Orbis Studio header, click your profile icon and select Settings.
  2. Click Model Providers.

Providers are grouped into Language Models and Embedding Models so you can see, at a glance, which models are available for generation and which are available for retrieval and indexing.

Connect a provider​

Most providers (such as OpenAI, Anthropic, or Azure) require an API key.

  1. Find the provider in the list.
  2. Enter the provider's API key and confirm the connection.

Once connected, the provider's models become selectable in model and embedding components across your flows.

info

Some providers run locally and do not require an API key. For example, Ollama is enabled by connecting to its endpoint rather than by entering a key.

Enable or disable models​

Each provider exposes an enabled/disabled state for its models. Disable the models you do not want builders to use, and enable the ones you do. This keeps the model picker focused on the options your organization actually supports.

Disconnect or deactivate a provider​

You can disconnect a provider's API key or deactivate the provider entirely.

warning

Disconnecting an API key or deactivating a provider disables all of that provider's models that are currently used in a flow. Any flow that depends on those models will stop working until you reconnect the provider or switch the affected components to another model.