Nice to see the project evolving beyond PostgreSQL. How does Databasus compare architecturally to tools like pgBackRest or WAL-G for Postgres, or to mysqldump-based solutions for MySQL? In particular, how do these differences affect performance and reliability for backups and restores?

Databasus and WAL-G have completely different architectures. WAL-G is installed on the server with DB. Databasus connects to the database remotely.

WAL-G is about PITR backups, Databasus is about logical ones.

From one side, Databasus usually is slower when database is large.

From another side, it's much easier to use (especially when you have a lot of DBs).

In the future, I am considering implementing incremental backups and following a similar approach as WAL-G or pgBackRest do. But it will be one of options, not mandatory one

Wanted something simple like this, but, this seems to place a hard requirement that database to import to should be empty before a restore.

Yes

To restore database - it should be empty (otherwise conflicts are possible)

Technically, you can try to restore to filled DB. But don't think it makes sense to try

Hmmm... can you not provide such a tool built on top of pgBackRest?

Im using it personally one one of my postgres installations, and it's incredibly simple.

Having web ui built on top would also make it accessible to a lot more people since setting it up the first time is not that trivial.

does this do point-in-time recovery? tried like 3 postgres backup tools, they all claim to support it but then you actually need to restore and it's a mess

No, it's logical backup (at least, by the time): PITR are complicated in setting up and recovering.