N/A

The common pgvector template on Railway (this one) includes postgres, so there's no need to have another postgres service as well

You would want to connect your easylanguagecomp service to the pgvector service using the pgvector service's private URL, which can be found in the environment variables page of that service.

If you create a reference variable in the easylanguagecomp service that references the pgvector's private URL, an arrow will appear on the canvas to visualize the link

Reading the template description for the pgvector template will give you some tips. https://railway.com/template/3jJFCA

Will read that first, thank you Adam

Is it possible that the DATABASE_URL_PRIVATE raises an error but DATABASE_URL works? Is there a difference other than the port?

I get "[Errno -2] Name or service not known" when using the PRIVATE one

There are some large differences between the public and private URL. The private URL can only be used with the service deployed on Railway, not locally. It utilizes the private network rather than the public network, which removes egress fees. The public URL is the publically exposed URL. this can be used locally

Here are the docs https://docs.railway.com/reference/private-networking

Thank you. Where can I run SQL to test if it is connected?

I completed the migrations but the "data" interface does not appear in PGvector as it did in postgres.

connect to the database with a database client like dbgate https://dbgate.org/download-community/

The data tab isn't very reliable, I never use it

I guess PGadmin would work too?

Yep, that would work

Thank you Adam

No problem, marking as solved now. If you have any further questions, please open a new thread

!s