Supabase Auth now supports Anonymous Sign-ins

Supabase Auth now supports anonymous sign-ins, one of our most-requested features by the community.

⚡️ Learn more about GA Week

Anonymous sign-ins can be used to create temporary users that haven’t signed up for your application yet. This lowers the friction for new users to try out your product since they don’t have to provide any signup credentials.

Enabling Anonymous sign-ins

You can enable anonymous sign-ins for your project today from the dashboard:

For local development, upgrade your Supabase CLI and add the config to the config.toml file:

[auth]
enable_anonymous_sign_ins = true 

You can create an anonymous user through the Javascript, Flutter or Swift SDKs today. Here’s how you can create an anonymous user using supabase-js.

const { data, error } = await supabase .auth .signInAnonymously()

Terminology

Profiles created with anonymous sign-ins are also authenticated!

Once you call .signInAnonymously() you have moved the user into an authentication flow, and we treat them like a signed-in user:

Restricting access for anonymous users

Like a permanent user, anonymous users are persisted in the auth.users table:

An anonymous user can be identified by the is_anonymous claim returned in the user’s JWT, which is accessible from your Row Level Security policies (RLS). This is helpful if you want to limit access to certain features in your application.

For example, let’s say that we have an online forum where users can create and read posts.

Given this table to store the posts:

create table public.posts ( id serial primary key, name text not null, description text
);

If we only want to allow permanent users to create posts, we can check if the user is anonymous by inspecting the JWT select auth.jwt() ->> 'is_anonymous'.

Using this function in an RLS policy:

create policy "Only permanent users can create posts"
on public.posts
for insert
to authenticated -- Note: user is still authenticated
with check ( (select auth.jwt() ->> 'is_anonymous')::boolean is false
); 

RLS gives us full flexibility to create a variety of rules.

For example, to allow read access for permanent users for all posts and limit anonymous users to posts created today:

create policy "Limit access to anonymous users"
on public.posts
for select
to authenticated -- Note: user is still authenticated
using ( case when (select (auth.jwt() ->> 'is_anonymous'))::boolean is true then (created_at >= current_date) else true end
);

Convert an anonymous user to a permanent user

At some point, an anonymous user may decide they want to create a post. This is where we prompt them to sign up for an account which converts them to a permanent user.

An anonymous user is considered a permanent user when they have an identity associated with them.
After they have been converted, the user id remains the same, which means that any data associated with the user’s id would be carried over.

Supabase Auth provides 2 ways to achieve this:

1- Link an email or phone identity
2- Link an OAuth identity

Link an email or phone identity

To link an email or phone identity:

const { data, error } = await supabase .auth .updateUser({ email })

Link an OAuth identity

To link an OAuth identity to an anonymous user, you need to enable manual linking for your project. Learn about how identity linkingworks with Supabase Auth.

Once enabled, you can call the linkIdentity() method:

const { data, error } = await supabase .auth .linkIdentity({ provider: 'google' })

Impersonating an anonymous user

When creating RLS policies to differentiate access for an anonymous user, you can leverage the user impersonation feature in the SQL editor to test out your policies:

Database role settings in the SQL editor. You can impersonate an anonymous user by selecting the user from the dropdown.

The user management screen provides an option to filter by anonymous users, which can help know how many anonymous users have been created.

What’s next

Managing anonymous users can be tricky, especially when you have a lot of visitors to your site. We’re working on an “automatic clean-up” option to delete anonymous users that have been inactive for more than 30 days. In the meantime, since anonymous users are stored in the auth schema in your database, you can clean up orphaned anonymous users by running the following query:

-- deletes anonymous users created more than 30 days ago
delete from auth.users
where is_anonymous is true and created_at < now() - interval '30 days'; 

We are also working on a linter to check your RLS policies and highlight those that allow anonymous users access – stay tuned for updates later this month!

Getting started

More on GA week