The open-source AI chat app for everyone organizations.
This is a fork of the excellent Chatbot-UI.
- Use OAuth/OIDC authentication provider, as declared in the
AUTH_PROVIDER
env var - Removed email/password signup/login page
- Pre-populate profile details from raw user auth data
- Allow profile name to be readonly
- Skip API keys in setup - relying on pre-configured env var keys instead
- Obtain
project_url
&service_role_key
from theVault
instead of hardcoding - Add an explicit
/logout
route - Check 'x-forwarded-*' headers for original public URL
- Name
- Database password (will be required later when linking to the project)
- Region
- Disable email provider
- Enable and configure an OAuth provider (can do this later if necessary)
supabase login
Obtain your Supabase project Reference ID
from the Project Settings
> General
page.
supabase link --project-ref <project-id>
You'll be asked for the database password you set when creating the project.
Now push the database schema...
supabase db push
In the Supabase UI...
From Project Settings
> API
...
- Copy
Project URL
& theservice_role
API key.
In Project Settings
> Vault
page, using Add new secret
create the secrets:
project_url
service_role_key
using the values copied from the API
page above.
(You may also need to create an encryption key the first you create a secret)
TODO
Follow original instructions for now.
You'll also need to add an additional environment variable:
AUTH_PROVIDER
- with the identifier of the provider selected in Supabase (eg.google
)
Once deployed, you'll need to go back to Supabase UI and in Authentication
> URL Configuration
,
set Site URL
to match your Vercel app URL.
docker build -t chatbot-ui-org-docker .
and to run locally:
docker run -p 3000:3000 chatbot-ui-org-docker
TODO?