Microsoft identity platform access tokens

Access tokens enable clients to securely call protected web APIs. Access tokens are used by web APIs to perform authentication and authorization.

Per the OAuth specification, access tokens are opaque strings without a set format. Some identity providers (IDPs) use GUIDs and others use encrypted blobs. The format of the access token can depend on how the API that accepts the token is configured.

Custom APIs registered by developers on the Microsoft identity platform can choose from two different formats of JSON Web Tokens (JWTs) called v1.0 and v2.0. Microsoft-developed APIs like Microsoft Graph or APIs in Azure have other proprietary token formats. These proprietary formats might be encrypted tokens, JWTs, or special JWT-like tokens that won't validate.

Clients must treat access tokens as opaque strings because the contents of the token are intended for the API only. For validation and debugging purposes only, developers can decode JWTs using a site like jwt.ms. Tokens that are received for a Microsoft API might not always be a JWT and can't always be decoded.

For details on what's inside the access token, clients should use the token response data that's returned with the access token to the client. When the client requests an access token, the Microsoft identity platform also returns some metadata about the access token for the consumption of the application. This information includes the expiry time of the access token and the scopes for which it's valid. This data allows the application to do intelligent caching of access tokens without having to parse the access token itself.

See the following sections to learn how an API can validate and use the claims inside an access token.

Note

All documentation on this page, except where noted, applies only to tokens issued for registered APIs. It doesn't apply to tokens issued for Microsoft-owned APIs, nor can those tokens be used to validate how the Microsoft identity platform issues tokens for a registered API.

Token formats

There are two versions of access tokens available in the Microsoft identity platform: v1.0 and v2.0. These versions determine the claims that are in the token and make sure that a web API can control the contents of the token.

Web APIs have one of the following versions selected as a default during registration:

  • v1.0 for Azure AD-only applications. The following example shows a v1.0 token (this token example won't validate because the keys have rotated prior to publication and personal information has been removed):

    eyJ0eXAiOiJKV1QiLCJhbGciOiJSUzI1NiIsIng1dCI6Imk2bEdrM0ZaenhSY1ViMkMzbkVRN3N5SEpsWSIsImtpZCI6Imk2bEdrM0ZaenhSY1ViMkMzbkVRN3N5SEpsWSJ9.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.D3H6pMUtQnoJAGq6AHd
    
  • v2.0 for applications that support consumer accounts. The following example shows a v1.0 token (this token example won't validate because the keys have rotated prior to publication and personal information has been removed):

    eyJ0eXAiOiJKV1QiLCJhbGciOiJSUzI1NiIsImtpZCI6Imk2bEdrM0ZaenhSY1ViMkMzbkVRN3N5SEpsWSJ9.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.pj4N-w_3Us9DrBLfpCt
    

The version can be set for applications by providing the appropriate value to the accessTokenAcceptedVersion setting in the app manifest. The values of null and 1 result in v1.0 tokens, and the value of 2 results in v2.0 tokens.

Token ownership

Two parties are involved in an access token request: the client, who requests the token, and the resource (Web API) that accepts the token. The aud claim in a token indicates the resource that the token is intended for (its audience). Clients use the token but shouldn't understand or attempt to parse it. Resources accept the token.

The Microsoft identity platform supports issuing any token version from any version endpoint - they aren't related. When accessTokenAcceptedVersion is set to 2, a client calling the v1.0 endpoint to get a token for that resource receives a v2.0 access token.

Resources always own their tokens using the aud claim and are the only applications that can change their token details.

Claims in access tokens

JWTs are split into three pieces:

  • Header - Provides information about how to validate the token including information about the type of token and how it was signed.
  • Payload - Contains all of the important data about the user or application that's attempting to call the service.
  • Signature - Is the raw material used to validate the token.

Each piece is separated by a period (.) and separately Base64 encoded.

Claims are present only if a value exists to fill it. An application shouldn't take a dependency on a claim being present. Examples include pwd_exp (not every tenant requires passwords to expire) and family_name (client credential flows are on behalf of applications that don't have names). Claims used for access token validation are always present.

Some claims are used to help the Microsoft identity platform secure tokens for reuse. These claims are marked as not being for public consumption in the description as Opaque. These claims may or may not appear in a token, and new ones may be added without notice.

Header claims

Claim Format Description
typ String - always JWT Indicates that the token is a JWT.
alg String Indicates the algorithm that was used to sign the token, for example, RS256.
kid String Specifies the thumbprint for the public key that can be used to validate this signature of the token. Emitted in both v1.0 and v2.0 access tokens.
x5t String Functions the same (in use and value) as kid. x5t and is a legacy claim emitted only in v1.0 access tokens for compatibility purposes.

Payload claims

Claim Format Description
aud String, an Application ID URI or GUID Identifies the intended audience of the token. The API must validate this value and reject the token if the value doesn't match. In v2.0 tokens, this value is always the client ID of the API. In v1.0 tokens, it can be the client ID or the resource URI used in the request. The value can depend on how the client requested the token.
iss String, a security token service (STS) URI Identifies the STS that constructs and returns the token, and the Azure AD tenant in which the user was authenticated. If the token issued is a v2.0 token (see the ver claim), the URI ends in /v2.0. The GUID that indicates that the user is a consumer user from a Microsoft account is 9188040d-6c67-4c5b-b112-36a304b66dad. The application can use the GUID portion of the claim to restrict the set of tenants that can sign in to the application, if applicable.
idp String, usually an STS URI Records the identity provider that authenticated the subject of the token. This value is identical to the value of the Issuer claim unless the user account isn't in the same tenant as the issuer, such as guests. If the claim isn't present, the value of iss can be used instead. For personal accounts being used in an organizational context (for instance, a personal account invited to an Azure AD tenant), the idp claim may be 'live.com' or an STS URI containing the Microsoft account tenant 9188040d-6c67-4c5b-b112-36a304b66dad.
iat int, a Unix timestamp Specifies when the authentication for this token occurred.
nbf int, a Unix timestamp Specifies the time before which the JWT must not be accepted for processing.
exp int, a Unix timestamp Specifies the expiration time on or after which the JWT must not be accepted for processing. A resource may reject the token before this time as well. The rejection can occur when a change in authentication is required or a token revocation has been detected.
aio Opaque String An internal claim used by Azure AD to record data for token reuse. Resources shouldn't use this claim.
acr String, a 0 or 1, only present in v1.0 tokens A value of 0 for the "Authentication context class" claim indicates the end-user authentication didn't meet the requirements of ISO/IEC 29115.
amr JSON array of strings, only present in v1.0 tokens Identifies how the subject of the token was authenticated.
appid String, a GUID, only present in v1.0 tokens The application ID of the client using the token. The application can act as itself or on behalf of a user. The application ID typically represents an application object, but it can also represent a service principal object in Azure AD.
azp String, a GUID, only present in v2.0 tokens A replacement for appid. The application ID of the client using the token. The application can act as itself or on behalf of a user. The application ID typically represents an application object, but it can also represent a service principal object in Azure AD.
appidacr String, a 0, 1, or 2, only present in v1.0 tokens Indicates how the client was authenticated. For a public client, the value is 0. If client ID and client secret are used, the value is 1. If a client certificate was used for authentication, the value is 2.
azpacr String, a 0, 1, or 2, only present in v2.0 tokens A replacement for appidacr. Indicates how the client was authenticated. For a public client, the value is 0. If client ID and client secret are used, the value is 1. If a client certificate was used for authentication, the value is 2.
preferred_username String, only present in v2.0 tokens. The primary username that represents the user. The value could be an email address, phone number, or a generic username without a specified format. The value is mutable and might change over time. Since the value is mutable, it must not be used to make authorization decisions. The value can be used for username hints, however, and in human-readable UI as a username. The profile scope is required in order to receive this claim.
name String Provides a human-readable value that identifies the subject of the token. The value isn't guaranteed to be unique, it's mutable, and is only used for display purposes. The profile scope is required in order to receive this claim.
scp String, a space separated list of scopes The set of scopes exposed by the application for which the client application has requested (and received) consent. The application should verify that these scopes are valid ones exposed by the application, and make authorization decisions based on the value of these scopes. Only included for user tokens.
roles Array of strings, a list of permissions The set of permissions exposed by the application that the requesting application or user has been given permission to call. For application tokens, this set of permissions is used during the client credential flow (v1.0, v2.0) in place of user scopes. For user tokens, this set of values is populated with the roles the user was assigned to on the target application.
wids Array of RoleTemplateID GUIDs Denotes the tenant-wide roles assigned to this user, from the section of roles present in Azure AD built-in roles. This claim is configured on a per-application basis, through the groupMembershipClaims property of the application manifest. Setting it to All or DirectoryRole is required. May not be present in tokens obtained through the implicit flow due to token length concerns.
groups JSON array of GUIDs Provides object IDs that represent the group memberships of the subject. These values are unique and can be safely used for managing access, such as enforcing authorization to access a resource. The groups included in the groups claim are configured on a per-application basis, through the groupMembershipClaims property of the application manifest. A value of null excludes all groups, a value of SecurityGroup includes only Active Directory Security Group memberships, and a value of All includes both Security Groups and Microsoft 365 Distribution Lists.

See the hasgroups claim for details on using the groups claim with the implicit grant. For other flows, if the number of groups the user is in goes over 150 for SAML and 200 for JWT, then Azure AD adds an overage claim to the claim sources. The claim sources point to the Microsoft Graph endpoint that contains the list of groups for the user.
hasgroups Boolean If present, always true, indicates whether the user is in at least one group. Used in place of the groups claim for JWTs in implicit grant flows if the full groups claim would extend the URI fragment beyond the URL length limits (currently six or more groups). Indicates that the client should use the Microsoft Graph API to determine the groups (https://graph.microsoft.com/v1.0/users/{userID}/getMemberObjects) of the user.
groups:src1 JSON object For token requests that aren't length limited (see hasgroups) but still too large for the token, a link to the full groups list for the user is included. For JWTs as a distributed claim, for SAML as a new claim in place of the groups claim.

Example JWT Value:
"groups":"src1"
"_claim_sources: "src1" : { "endpoint" : "https://graph.microsoft.com/v1.0/users/{userID}/getMemberObjects" }
sub String The principal about which the token asserts information, such as the user of an application. This value is immutable and can't be reassigned or reused. It can be used to perform authorization checks safely, such as when the token is used to access a resource, and can be used as a key in database tables. Because the subject is always present in the tokens that Azure AD issues, use this value in a general-purpose authorization system. The subject is, however, a pairwise identifier that is unique to a particular application ID. If a single user signs into two different applications using two different client IDs, those applications receive two different values for the subject claim. Two different values may or may not be desired depending on architecture and privacy requirements. See also the oid claim (which does remain the same across applications within a tenant).
oid String, a GUID The immutable identifier for the requestor, which is the user or service principal whose identity has been verified. It can also be used to perform authorization checks safely and as a key in database tables. This ID uniquely identifies the requestor across applications. Two different applications signing in the same user receive the same value in the oid claim. The oid can be used when making queries to Microsoft online services, such as the Microsoft Graph. The Microsoft Graph returns this ID as the id property for a given user account. Because the oid allows multiple applications to correlate principals, the profile scope is required in order to receive this claim for users. If a single user exists in multiple tenants, the user contains a different object ID in each tenant. The accounts are considered different, even though the user logs into each account with the same credentials.
tid String, a GUID Represents the tenant that the user is signing in to. For work and school accounts, the GUID is the immutable tenant ID of the organization that the user is signing in to. For sign-ins to the personal Microsoft account tenant (services like Xbox, Teams for Life, or Outlook), the value is 9188040d-6c67-4c5b-b112-36a304b66dad. To receive this claim, the application must request the profile scope.
unique_name String, only present in v1.0 tokens Provides a human readable value that identifies the subject of the token. This value isn't guaranteed to be unique within a tenant and should be used only for display purposes.
uti String Token identifier claim, equivalent to jti in the JWT specification. Unique, per-token identifier that is case-sensitive.
rh Opaque String An internal claim used by Azure to revalidate tokens. Resources shouldn't use this claim.
ver String, either 1.0 or 2.0 Indicates the version of the access token.

Groups overage claim

Azure AD limits the number of object IDs that it includes in the groups claim to stay within the size limit of the HTTP header. If a user is a member of more groups than the overage limit (150 for SAML tokens, 200 for JWT tokens, and only 6 if issued by using the implicit flow), then Azure AD doesn't emit the groups claim in the token. Instead, it includes an overage claim in the token that indicates to the application to query the Microsoft Graph API to retrieve the group membership of the user.

{
    ...
    "_claim_names": {
        "groups": "src1"
    },
    "_claim_sources": {
        "src1": {
            "endpoint": "[Url to get this user's group membership from]"
        }   
    }
    ...
}

Use the BulkCreateGroups.ps1 provided in the App Creation Scripts folder to help test overage scenarios.

v1.0 basic claims

The following claims are included in v1.0 tokens if applicable, but aren't included in v2.0 tokens by default. To use these claims for v2.0, the application requests them using optional claims.

Claim Format Description
ipaddr String The IP address the user authenticated from.
onprem_sid String, in SID format In cases where the user has an on-premises authentication, this claim provides their SID. Use this claim for authorization in legacy applications.
pwd_exp int, a Unix timestamp Indicates when the user's password expires.
pwd_url String A URL where users can be sent to reset their password.
in_corp boolean Signals if the client is signing in from the corporate network. If they aren't, the claim isn't included.
nickname String Another name for the user, separate from first or last name.
family_name String Provides the last name, surname, or family name of the user as defined on the user object.
given_name String Provides the first or given name of the user, as set on the user object.
upn String The username of the user. May be a phone number, email address, or unformatted string. Should only be used for display purposes and providing username hints in reauthentication scenarios.

amr claim

Identities can authenticate in different ways, which may be relevant to the application. The amr claim is an array that can contain multiple items, such as ["mfa", "rsa", "pwd"], for an authentication that used both a password and the Authenticator app.

Value Description
pwd Password authentication, either a user's Microsoft password or a client secret of an application.
rsa Authentication was based on the proof of an RSA key, for example with the Microsoft Authenticator app. This value also indicates whether authentication was done by a self-signed JWT with a service owned X509 certificate.
otp One-time passcode using an email or a text message.
fed A federated authentication assertion (such as JWT or SAML) was used.
wia Windows Integrated Authentication
mfa Multi-factor authentication was used. When this claim is present, the other authentication methods are included.
ngcmfa Equivalent to mfa, used for provisioning of certain advanced credential types.
wiaormfa The user used Windows or an MFA credential to authenticate.
none No authentication was done.

Access token lifetime

The default lifetime of an access token is variable. When issued, the default lifetime of an access token is assigned a random value ranging between 60-90 minutes (75 minutes on average). The variation improves service resilience by spreading access token demand over a time, which prevents hourly spikes in traffic to Azure AD.

Tenants that don’t use Conditional Access have a default access token lifetime of two hours for clients such as Microsoft Teams and Microsoft 365.

The lifetime of an access token can be adjusted to control how often the client application expires the application session, and how often it requires the user to reauthenticate (either silently or interactively). To override the default access token lifetime variation, set a static default access token lifetime by using Configurable token lifetime (CTL).

Default token lifetime variation is applied to organizations that have Continuous Access Evaluation (CAE) enabled. Default token lifetime variation is applied even if the organizations use CTL policies. The default token lifetime for long lived token lifetime ranges from 20 to 28 hours. When the access token expires, the client must use the refresh token to silently acquire a new refresh token and access token.

Organizations that use Conditional Access sign-in frequency (SIF) to enforce how frequently sign-ins occur can't override default access token lifetime variation. When organizations use SIF, the time between credential prompts for a client is the token lifetime that ranges from 60 - 90 minutes plus the sign-in frequency interval.

Here's an example of how default token lifetime variation works with sign-in frequency. Let's say an organization sets sign-in frequency to occur every hour. The actual sign-in interval occurs anywhere between 1 hour to 2.5 hours because the token is issued with lifetime ranging from 60-90 minutes (due to token lifetime variation).

If a user with a token with a one hour lifetime performs an interactive sign-in at 59 minutes (just before the sign-in frequency being exceeded), there's no credential prompt because the sign-in is below the SIF threshold. If a new token is issued with a lifetime of 90 minutes, the user wouldn't see a credential prompt for another hour and a half. When a silent renewal attempted of the 90-minute token lifetime is made, Azure AD requires a credential prompt because the total session length has exceeded the sign-in frequency setting of 1 hour. In this example, the time difference between credential prompts due to the SIF interval and token lifetime variation would be 2.5 hours.

Validate tokens

Not all applications should validate tokens. Only in specific scenarios should applications validate a token:

  • Web APIs must validate access tokens sent to them by a client. They must only accept tokens containing their aud claim.
  • Confidential web applications like ASP.NET Core must validate ID tokens sent to them by using the user's browser in the hybrid flow, before allowing access to a user's data or establishing a session.

If none of the above scenarios apply, the application won't benefit from validating the token, and may present a security and reliability risk if decisions are made based on the validity of the token. Public clients like native or single-page applications don't benefit from validating tokens because the application communicates directly with the IDP where SSL protection ensures the tokens are valid.

APIs and web applications must only validate tokens that have an aud claim that matches the application. Other resources may have custom token validation rules. For example, tokens for Microsoft Graph won't validate according to these rules due to their proprietary format. Validating and accepting tokens meant for another resource is an example of the confused deputy problem.

If the application needs to validate an ID token or an access token, it should first validate the signature of the token and the issuer against the values in the OpenID discovery document. For example, the tenant-independent version of the document is located at https://login.microsoftonline.com/common/.well-known/openid-configuration.

The Azure AD middleware has built-in capabilities for validating access tokens, see samples to find one in the appropriate language. There are also several third-party open-source libraries available for JWT validation. For more information about Azure AD authentication libraries and code samples, see the authentication libraries.

Validating the signature

A JWT contains three segments, which are separated by the . character. The first segment is known as the header, the second as the body, and the third as the signature. The signature segment can be used to validate the authenticity of the token so that it can be trusted by the application.

Tokens issued by Azure AD are signed using industry standard asymmetric encryption algorithms, such as RS256. The header of the JWT contains information about the key and encryption method used to sign the token:

{
  "typ": "JWT",
  "alg": "RS256",
  "x5t": "iBjL1Rcqzhiy4fpxIxdZqohM2Yk",
  "kid": "iBjL1Rcqzhiy4fpxIxdZqohM2Yk"
}

The alg claim indicates the algorithm that was used to sign the token, while the kid claim indicates the particular public key that was used to validate the token.

At any given point in time, Azure AD may sign an ID token using any one of a certain set of public-private key pairs. Azure AD rotates the possible set of keys on a periodic basis, so the application should be written to handle those key changes automatically. A reasonable frequency to check for updates to the public keys used by Azure AD is every 24 hours.

Acquire the signing key data necessary to validate the signature by using the OpenID Connect metadata document located at:

https://login.microsoftonline.com/common/v2.0/.well-known/openid-configuration

Tip

Try this in a browser: URL

The following information describes the metadata document:

  • Is a JSON object that contains several useful pieces of information, such as the location of the various endpoints required for doing OpenID Connect authentication.
  • Includes a jwks_uri, which gives the location of the set of public keys that correspond to the private keys used to sign tokens. The JSON Web Key (JWK) located at the jwks_uri contains all of the public key information in use at that particular moment in time. The JWK format is described in RFC 7517. The application can use the kid claim in the JWT header to select the public key, from this document, which corresponds to the private key that has been used to sign a particular token. It can then do signature validation using the correct public key and the indicated algorithm.

Note

Use the kid claim to validate the token. Though v1.0 tokens contain both the x5t and kid claims, v2.0 tokens contain only the kid claim.

Doing signature validation is outside the scope of this document. There are many open-source libraries available for helping with signature validation if necessary. However, the Microsoft identity platform has one token signing extension to the standards, which are custom signing keys.

If the application has custom signing keys as a result of using the claims-mapping feature, append an appid query parameter that contains the application ID to get a jwks_uri that points to the signing key information of the application, which should be used for validation. For example: https://login.microsoftonline.com/{tenant}/.well-known/openid-configuration?appid=6731de76-14a6-49ae-97bc-6eba6914391e contains a jwks_uri of https://login.microsoftonline.com/{tenant}/discovery/keys?appid=6731de76-14a6-49ae-97bc-6eba6914391e.

Claims based authorization

The business logic of the application dictates claims based authorization. Some common authorization methods are listed below.

Validate the token

Use the aud claim to ensure that the user intended to call the application. If the identifier of the resource isn't in the aud claim, reject it.

Validate user permission

Use the roles and wids claims to validate that the user has authorization to call the API. For example, an administrator may have permission to write to the API, but not a normal user. Check that the tid inside the token matches the tenant ID used to store the data in the API.

When a user stores data in the API from one tenant, they must sign into that tenant again to access that data. Never allow data in one tenant to be accessed from another tenant.

Use the amr claim to verify the user has performed MFA. The enforcement of MFA is done using Conditional Access. If roles or groups claims are requested in the access token, verify that the user is in the group allowed to do this action.

For tokens retrieved using the implicit flow, query the Microsoft Graph for this data, as it's often too large to fit in the token.

Don't use email or upn claim values to determine whether the user in an access token should have access to data. Mutable claim values like these can change over time, making them insecure and unreliable for authorization.

Use immutable claim values tid and sub or oid as a combined key for uniquely identifying the API's data and determining whether a user should be granted access to that data.

  • Good: tid + sub
  • Better: tid + oid - the oid is consistent across applications, so an ecosystem of applications can audit user access to data.

Don't use mutable, human-readable identifiers like email or upn for uniquely identifying data.

Validate application sign-in

  • Use the scp claim to validate that the user has granted the calling app permission to call your API.
  • Ensure the calling client is allowed to call your API using the appid claim (for v1.0 tokens) or the azp claim (for v2.0 tokens).
    • You only need to validate these claims (appid, azp) if you want to restrict your web API to be called only by pre-determined applications (e.g., line-of-business applications or web APIs called by well-known frontends). APIs intended to allow access from any calling application do not need to validate these claims.

User and application tokens

An application may receive tokens for a user or directly from an application through the client credentials flow. These app-only tokens indicate that this call is coming from an application and doesn't have a user backing it. These tokens are handled largely the same:

  • Use roles to see permissions that have been granted to the subject of the token.
  • Use oid or sub to validate that the calling service principal is the expected one.

If the application needs to distinguish between app-only access tokens and access tokens for users, use the idtyp optional claim. To detect app-only access tokens, add the idtyp claim to the accessToken field, and check for the value app. ID tokens and access tokens for users won't have the idtyp claim included.

Token revocation

Refresh tokens can be invalidated or revoked at any time, for different reasons. The reasons fall into the categories of timeouts and revocations.

Token timeouts

When an organization uses token lifetime configuration, the lifetime of refresh tokens can be altered. It's expected that some tokens can go without use. For example, the user doesn't open the application for three months and then the token expires. Applications can encounter scenarios where the login server rejects a refresh token due to its age.

  • MaxInactiveTime: If the refresh token hasn't been used within the time dictated by the MaxInactiveTime, the refresh token is no longer valid.
  • MaxSessionAge: If MaxAgeSessionMultiFactor or MaxAgeSessionSingleFactor have been set to something other than their default (Until-revoked), then reauthentication is required after the time set in the MaxAgeSession* elapses. Examples:
    • The tenant has a MaxInactiveTime of five days, and the user went on vacation for a week, and so Azure AD hasn't seen a new token request from the user in seven days. The next time the user requests a new token, they'll find their refresh token has been revoked, and they must enter their credentials again.
    • A sensitive application has a MaxAgeSessionSingleFactor of one day. If a user logs in on Monday, and on Tuesday (after 25 hours have elapsed), they'll be required to reauthenticate.

Token revocations

Refresh tokens can be revoked by the server due to a change in credentials, or due to use or administrative action. Refresh tokens are in the classes of confidential clients and public clients.

Change Password-based cookie Password-based token Non-password-based cookie Non-password-based token Confidential client token
Password expires Stays alive Stays alive Stays alive Stays alive Stays alive
Password changed by user Revoked Revoked Stays alive Stays alive Stays alive
User does SSPR Revoked Revoked Stays alive Stays alive Stays alive
Admin resets password Revoked Revoked Stays alive Stays alive Stays alive
User revokes their refresh tokens by using PowerShell Revoked Revoked Revoked Revoked Revoked
Admin revokes all refresh tokens for a user by using PowerShell Revoked Revoked Revoked Revoked Revoked
Single sign-out (v1.0, v2.0 ) on web Revoked Stays alive Revoked Stays alive Stays alive

Non-password-based

A non-password-based login is one where the user didn't type in a password to get it. Examples of non-password-based login include:

  • Using your face with Windows Hello
  • FIDO2 key
  • SMS
  • Voice
  • PIN

Check out Primary Refresh Tokens for more details on primary refresh tokens.

Next steps