The first version of Dolt mutated data at rest in file systems. Users dolt clone databases, and run Git commands like dolt checkout, dolt branch, dolt merge, ...etc.

Dolt is increasingly used as a server to accommodate UX and scaling challenges. When getting started with Dolt, however, the cli is easiest way to experiment with editing databases.

To compare, here is a file-system getting started:

$ sudo bash -c 'curl -L | sudo bash'
$ mkdir my-database
$ cd my-database
$ dolt init
$ dolt sql -q "create table t1 (a int primary key, b int)"
$ dolt ls

The server requires every step above, including a server instantiation:

$ dolt sql-server -l trace --max-connections 10

and a separately downloaded client to interact with our server:

$ mysql --user=root --host= -p my_database
Enter password:
mysql> show tables;
| Table |
| t1 |
1 row in set (0.00 sec)

Any tool that works with MySQL works equivalently with a Dolt server. The downside is you need two tools: dolt for the server, something else for the client.

Edit on GitHub