Importing a workspace used to have a quiet failure mode. The tables arrived, the flows arrived, and then the flow that sends email did nothing, because the SMTP connection it was built against lived in the account you exported from. You found out days later, from the customer who never got their confirmation.
Import now asks first.
It tells you what it needs
An archive records the connections its automations depend on: the SMTP server a flow sends through, the OpenAI key behind an AI step, the API connection an API Request was built against, and the mail server a table uses for record email.
When you import, a Requirements section lists each one and what uses it. Point it at one of your own connections, or click Set up a new one to create it in a new tab and come back. Where the archive can tell which provider a connection points at, you get that provider's own setup screen with its endpoints already filled in, rather than a blank form.
The Import button stays disabled until every requirement is answered. That is on purpose. A half-connected import produces automations that look installed and quietly fail, which is worse than one that refuses to start.
Credentials never travel. Only the name and type of each connection are recorded, so the import knows what to ask you for.
One archive, several workspaces
Export is no longer one workspace at a time. Pick as many as you like, expand each to choose its tables, and it all comes out as a single .zip. On the way back in you choose a destination per workspace: a new one, or merge into something you already have.
Links between tables in different workspaces survive the trip. Records, files and comments come across for every workspace in the archive, not just the first.
The same applies to the App Market. A template that spans several workspaces now installs as several workspaces instead of being flattened into one, and its public page lists the connections you will need before you click Install.
A plain SQL dump, if you want one
There is a new Raw SQL dump toggle on export. Tick it and the archive also contains a CREATE TABLE and INSERT file per workspace, the kind of thing you can feed straight to mysql.
Untick everything else and that is nearly all you get: a small archive with your table definitions and your data, ready to load into your own database or hand to whoever asked for it. It is there for use outside InfoLobby. Importing an archive back into InfoLobby ignores the dump and rebuilds properly from the structured files, so your fields, views, reports and flows come back with it.
Full details in the export and import guide.