Skip to main content


Using an Auth.js / NextAuth.js adapter you can connect to any database service or even several different services at the same time. The following listed official adapters are created and maintained by the community:


If you don't find an adapter for the database or service you use, you can always create one yourself. Have a look at our guide on how to create a database adapter.


Auth.js can be used with any database. Models tell you what structures Auth.js expects from your database. Models will vary slightly depending on which adapter you use, but in general, will look something like this:

More information about each Model/Table can be found below.


You can create your adapter if you want to use Auth.js with a database that is not supported out of the box, or you have to change fields on any of the models.


The User model is for information such as the user's name and email address.

Email address is optional, but if one is specified for a User, then it must be unique.


If a user first signs in with an OAuth provider, then their email address is automatically populated using the one from their OAuth profile if the OAuth provider returns one.

This provides a way to contact users and for users to maintain access to their account and sign in using email in the event they are unable to sign in with the OAuth provider in the future (if the Email Provider is configured).

User creation in the database is automatic and happens when the user is logging in for the first time with a provider. If the first sign-in is via the OAuth Provider, the default data saved is id, name, email and image. You can add more profile data by returning extra fields in your OAuth provider's profile() callback.

If the first sign-in is via the Email Provider, then the saved user will have id, email, emailVerified, where emailVerified is the timestamp of when the user was created.


The Account model is for information about OAuth accounts associated with a User

A single User can have multiple Accounts, but each Account can only have one User.

Account creation in the database is automatic and happens when the user is logging in for the first time with a provider, or the Adapter.linkAccount method is invoked. The default data saved is access_token, expires_at, refresh_token, id_token, token_type, scope and session_state. You can save other fields or remove the ones you don't need by returning them in the OAuth provider's account() callback.

Linking Accounts to Users happen automatically, only when they have the same e-mail address, and the user is currently signed in. Check the FAQ for more information on why this is a requirement.


You can manually unlink accounts if your adapter implements the unlinkAccount method. Make sure to take all the necessary security steps to avoid data loss.


Linking and unlinking accounts through an API is a planned feature:


The Session model is used for database sessions. It is not used if JSON Web Tokens are enabled. Keep in mind, that you can use a database to persist Users and Accounts, and still use JWT for sessions. See the session.strategy option.

A single User can have multiple Sessions, each Session can only have one User.


When a Session is read, we check if its expires field indicates an invalid session, and delete it from the database. You can also do this clean-up periodically in the background to avoid our extra delete call to the database during an active session retrieval. This might result in a slight performance increase in a few cases.

Verification Token

The Verification Token model is used to store tokens for passwordless sign in.

A single User can have multiple open Verification Tokens (e.g. to sign in to different devices).

It has been designed to be extendable for other verification purposes in the future (e.g. 2FA / magic codes, etc.).


Auth.js makes sure that every token is usable only once, and by default has a short (1 day, can be configured by maxAge) lifetime. If your user did not manage to finish the sign-in flow in time, they will have to start the sign-in process again.


Due to users forgetting or failing at the sign-in flow, you might end up with unwanted rows in your database, that you might have to periodically clean up to avoid filling the database up with unnecessary data.

RDBMS Naming Convention

Auth.js / NextAuth.js uses camelCase for its database rows while respecting the conventional snake_case formatting for OAuth-related values. If the mixed casing is an issue for you, most adapters have a dedicated documentation section on how to force a casing convention.


Check out the @auth/core/adapters API Reference documentation.