> ## Documentation Index
> Fetch the complete documentation index at: https://familyco.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Provider connections

> Connect AI providers, pick a primary, and select default models.

Provider connections let the Founder configure which AI providers and models FamilyCo uses. Connections persist API keys (or OAuth tokens) and a primary-provider selection.

## Views

* **Provider inventory** with connection status
* **Available models** and the current selected model
* **Auth type** (`apikey` / `oauth`) and primary-provider marker

## Actions

* Test a provider connection
* Connect / disconnect a provider with API key
* Connect a provider via OAuth *(desktop runtime only)*
* Select primary provider and default model

## Constraints

* A provider must be connected before it can become primary.
* OAuth endpoints are unavailable in **server** runtime mode (they require the desktop shell).
* Connection and selection actions produce audit records.

## Where defaults apply

* `Company.defaultProvider` and `Company.defaultModel` are the company-wide defaults.
* Each agent may override with its own `provider`, `defaultModel`, `temperature`, and `maxTokens`.
* The chat session uses the recipient agent's settings; the Executive Agent inherits company defaults unless overridden.

## Renderer environment

When the web client runs against an external server, it reads its own variables:

* `VITE_API_BASE_URL`
* `VITE_API_KEY`
* `VITE_BEARER_TOKEN`

See [Configuration](/guides/configuration) for the complete environment reference.