Authentication

b.well MCP Server offers multiple ways to authenticate:

1. Federated Authentication - Client passes their own token. No b.well login screen is shown to the user. b.well exchanges the client's token for a b.well token. The client then passes the b.well token to subsequent MCP calls. This method is ideal for applications that already have their own authentication system and want to leverage existing user credentials.

2. Dynamic Client Registration - Client uses DCR to dynamically create a client id and client secret with b.well. During the authentication flow, the user is shown a b.well login screen where they enter their b.well username and password. Once authenticated, a b.well token is issued and can be used for MCP calls. This approach is useful for applications that need to register programmatically without pre-configured credentials.

3. Static Client Registration - Client receives a client id and client secret from b.well through a manual registration process. The client enters these OAuth credentials and uses them to contact b.well's authentication endpoints. The user is shown a b.well login screen where they enter their b.well username and password. Upon successful authentication, a b.well token is issued for use with MCP calls. This method is suitable for applications with a stable, pre-registered identity.

Each authentication method serves different use cases depending on your application's architecture and requirements. Choose the method that best aligns with your existing authentication infrastructure and user experience goals.