Database
Better Auth requires a database connection to store data. It comes with a query builder called Kysley to manage and query your database. The database will be used to store data such as users, sessions, and more. Plugins can also define their own database tables to store data.
You can pass a database connection to Better Auth by passing a supported database instance, a dialect instance or a Kysely instance in the database options.
Example: Sqlite
Example: Postgres
Example: MySQL
Example: Custom Dialect using libSQL
See Kysely Dialects for more dialects supported by Kysely.
Example: Custom Kysely Instance
Using Adapters
If your database is managed by an ORM like Prisma or Drizzle, you can use the corresponding adapter to connect to the database. Better Auth comes with built-in adapters for Prisma and Drizzle. You can pass the adapter to the database
object in the auth options.
Prisma Adapter
The Prisma adapter expects a prisma client instance and a provider key that specifies the database provider to use. The provider key can be sqlite
, postgres
, mysql
, or any other supported by prisma.
Drizzle adapter
The Drizzle adapter expects a drizzle client instance and a provider key that specifies the database provider to use. The provider key can be sqlite
, pg
or mysql
.
Mapping Schema
The Drizzle adapter expects the schema you define to match the table names. For example, if your Drizzle schema maps the user
table to users
, you need to manually pass the schema and map it to the user table.
MongoDB Adapter
The MongoDB adapter expects a mongodb client instance and a database name. The adapter will create a new database with the provided name if it doesn't exist.
Use Plural Form
If your schema uses plural form for table names, you can pass the usePlural
option to the adapter.
CLI
Better Auth comes with a CLI tool to manage database migrations and generate schema.
Running Migrations
The cli checks your database and prompts you to add missing tables or update existing ones with new columns. This is only supported for the built-in Kysely adapter. For other adapters, you can use the generate
command to create the schema and handle the migration through your ORM.
Generating Schema
Better Auth also provides a generate
command to generate the schema required by Better Auth. The generate
command creates the schema required by Better Auth. If you're using a database adapter like Prisma or Drizzle, this command will generate the right schema for your ORM. If you're using the built-in Kysely adapter, it will generate an SQL file you can run directly on your database.
See the CLI documentation for more information on the CLI.
If you prefer adding tables manually, you can do that as well. The core schema required by Better Auth is described below and you can find additional schema required by plugins in the plugin documentation.
Secondary Storage
Secondary storage in Better Auth allows you to use key-value stores for managing session data, rate limiting counters, etc. This can be useful when you want to offload the storage of this intensive records to a high performance storage or even RAM.
Implementation
To use secondary storage, implement the SecondaryStorage
interface:
Then, provide your implementation to the betterAuth
function:
Example: Redis Implementation
Here's a basic example using Redis:
This implementation allows Better Auth to use Redis for storing session data and rate limiting counters.
Core Schema
Better Auth requires the following tables to be present in the database. The types are in typescript
format. You can use corresponding types in your database.
User
Table Name: user
Field Name | Type | Key | Description |
---|---|---|---|
id | string | Unique identifier for each user | |
name | string | - | User's chosen display name |
string | - | User's email address for communication and login | |
emailVerified | boolean | - | Whether the user's email is verified |
image | string | User's image url | |
createdAt | Date | - | Timestamp of when the user account was created |
updatedAt | Date | - | Timestamp of the last update to the user's information |
Session
Table Name: session
Field Name | Type | Key | Description |
---|---|---|---|
id | string | Unique identifier for each session. This is used as the session token. | |
userId | string | The id of the user | |
expiresAt | Date | - | The time when the session expires |
ipAddress | string | The IP address of the device | |
userAgent | string | The user agent information of the device |
Account
Table Name: account
Field Name | Type | Key | Description |
---|---|---|---|
id | string | Unique identifier for each account | |
userId | string | The id of the user | |
accountId | string | - | The id of the account as provided by the SSO or equal to userId for credential accounts |
providerId | string | - | The id of the provider |
accessToken | string | The access token of the account. Returned by the provider | |
refreshToken | string | The refresh token of the account. Returned by the provider | |
expiresAt | Date | The time when the access token expires | |
password | string | The password of the account. Mainly used for email and password authentication |
Verification
Table Name: verification
Field Name | Type | Key | Description |
---|---|---|---|
id | string | Unique identifier for each verification | |
identifier | string | - | The identifier for the verification request |
value | string | - | The value to be verified |
expiresAt | Date | - | The time when the verification request expires |
createdAt | Date | - | Timestamp of when the verification request was created |
Custom Tables
Better Auth allows you to customize the table names and column names for the core schema. You can also extend the core schema by adding additional fields to the user and session tables.
Custom Table Names
You can customize the table names and column names for the core schema by using the modelName
and fields
properties in your auth config:
Type inference in your code will still use the original field names (e.g., user.name
, not user.full_name
).
To customize table names and column name for plugins, you can use the schema
property in the plugin config:
Extending Core Schema
Better Auth provides a type-safe way to extend the user
and session
schemas. You can add custom fields to your auth config, and the CLI will automatically update the database schema. These additional fields will be properly inferred in functions like useSession
, signUp.email
, and other endpoints that work with user or session objects.
To add custom fields, use the additionalFields
property in the user
or session
object of your auth config. The additionalFields
object uses field names as keys, with each value being a FieldAttributes
object containing:
type
: The data type of the field (e.g., "string", "number", "boolean").required
: A boolean indicating if the field is mandatory.defaultValue
: The default value for the field (note: this only applies in the JavaScript layer; in the database, the field will be optional).input
: This determines whether a value can be provided when creating a new record (default:true
). If there are additional fields, likerole
, that should not be provided by the user during signup, you can set this tofalse
.
Here's an example of how to extend the user schema with additional fields:
Now you can access the additional fields in your application logic.
See the Typescript documentation for more information on how to infer additional fields on the client side.
ID Generation
Better Auth uses nanoId
by default to generate unique IDs for users, sessions, and other entities. If you want to customize how IDs are generated, you can pass your own function through the adapter or database configuration.
Guidelines:
- IDs need to be in string format.
- The ID should be unique for each entity type (like users, sessions, etc.).
- Since the session ID acts as the token, it should be both unique and sufficiently long for security.
Example: Built In Kysley Adapter
Example: Prisma Adapter
You can also disable ID generation by setting the generateId
option to false
in your adapter configuration. This will assume your database will generate the ID automatically.
Database Hooks
Database hooks allow you to define custom logic that can be executed during the lifecycle of core database operations in Better Auth. You can create hooks for the following models: user, session, and account.
There are two types of hooks you can define:
1. Before Hook
- Purpose: This hook is called before the respective entity (user, session, or account) is created or updated.
- Behavior: If the hook returns
false
, the operation will be aborted. And If it returns a data object, it'll replace the orginal payload.
2. After Hook
- Purpose: This hook is called after the respective entity is created or updated.
- Behavior: You can perform additional actions or modifications after the entity has been successfully created or updated.
Example Usage
Plugins Schema
Plugins can define their own tables in the database to store additional data. They can also add columns to the core tables to store additional data. For example, the two factor authentication plugin adds the following columns to the user
table:
twoFactorEnabled
: Whether two factor authentication is enabled for the user.twoFactorSecret
: The secret key used to generate TOTP codes.twoFactorBackupCodes
: Encrypted backup codes for account recovery.
To add new tables and columns to your database, you have two options:
CLI
: Use the migrate or generate command. These commands will scan your database and guide you through adding any missing tables or columns.
Manual Method
: Follow the instructions in the plugin documentation to manually add tables and columns.
Both methods ensure your database schema stays up-to-date with your plugins' requirements.