No. Emolysis does not bill you for Databricks compute, or for Snowflake or Microsoft Fabric compute once those platforms are supported. Deploys and production runs use your platform's compute and appear on your platform's bill, as they do today.
Sampling reads a slice of your data once, and that read is a query on your platform. After that, iterating runs on DuckDB on your machine and uses no cluster hours.
Pull request checks (emo check in your CI; the GitHub App is planned) include a cost estimate for your target platform on every pull request, so reviewers can see what a change is expected to cost before it merges. Read how local-first works and how deploys run.