Workflow overview

Model deployment starts on the left. Models can be deployed from Python notebooks like Jupyter and Hex using mb.deploy(...). Models can also be deployed using git push to Modelbit or through your organization's GitHub or GitLab repo.

Once deployed, you can call models from SQL functions defined in your Snowflake or Redshift warehouse. These SQL functions are natural interfaces for calling models from dbt or exploring models in Looker.

You can also call your models from a REST API to add inferences to your product, mobile app, or website.