FAQ and troubleshooting

“The AI service is unavailable right now. Please try again.”

The AI service couldn’t be reached. Generation still works with built-in values, and SQL/CSV imports fall back to the built-in parser. Try again in a few minutes.

”You’ve used all 50 AI actions for today”

You’ve reached the daily allowance. It refills over the next 24 hours; meanwhile, generation uses built-in values and cached AI values still work. See Limits and quotas.

I need more tables, rows or AI actions.

Enterprise plans raise these limits and can include other AI models. Email contact@teragenai.com — see Enterprise limits.

My import found the wrong types.

Fix them in the editor after clicking Use this schema — or add an example row to the file (e.g. an INSERT) so types and formats are detected from real-looking values.

Can I import an Erwin model?

Yes, but not the native .erwin file. Export the model from Erwin as XML or SQL DDL and upload that, or upload an image of the diagram. See Import an existing model.

Some columns came out as text in my database.

That’s deliberate when values don’t fit the database type (for example CUST-000001 in an ID column). Remove the sample value, or change the column type, to get a native type.

Which file should I run?

Always create_tables.sql first, then either the loader (load_files.sql / load_files.duckdb.sql) or insert_backup.sql. The zip’s README has the exact commands for your database.

Can I use the data in production?

It’s built for testing, development, demos and training. It’s realistic, but it’s not real.