Skip to main content
A credential holds the sensitive value used to authenticate to another system. A connection refers to that credential and stores the non-sensitive details of the source. Open Connections → Credentials to manage credentials. The platform also uses the term secret for these stored values.

​
Create a credential

Use New credential in the Credentials tab, or Create credential while setting up a connection. Select the type required by the source, enter its values, and use a descriptive name. Examples include username and password, client ID and secret, AWS access keys, an Azure storage account key, and a Google service account key. Google Sheets and LinkedIn use their supported authorisation flows. Paste a service account key only into its dedicated credential field. Do not include passwords, tokens, or private keys in chat, dataset descriptions, or shared context.

​
Reuse deliberately

A credential can be used by more than one workflow. Before replacing or removing it, identify the connections and jobs that depend on it. Use separate credentials when different data sources need different access or ownership. Saved values remain hidden in the connection interface. To replace access, create a new credential, select it in the affected connections, and verify a small run before revoking the old provider credential.

​
Delete and revoke are different

Removing a credential from Dawn and revoking it at the provider are separate operations. To prevent future access at the source, revoke or rotate it in the provider’s administration tools as well. Neither action removes copies of data already imported into Dawn. See workspace cleanup for that separate task, and permissions and credentials for provider access boundaries.