Using variables in dbt cloud

I’m working on orchestration of our dbt-cloud models using airflow. I’d like to make an incremental models updated for specific dates. The only way I’ve found to make it work is to provide a date as a variable. That works well for local dbt setup, but I don’t see a way to make it work in dbt cloud. I don’t see how I can supply a variable for a dbt run in the cloud.
The only way I’m currently see that may work is to create new model for each run and hardcode variable in execute_steps parameter. But that is an overkill and I don’t like that solution.
Do you know another workaround?