Sidemantic is a **semantic layer**: a system for defining **models**, **dimensions**, **metrics**, and **relationships** once, then querying them consistently.

## What You Get

- **A contract**: metrics have one definition (aggregation, filters, grain).
- **Automatic joins**: cross-model queries infer join paths from relationships (no `JOIN` keyword in semantic SQL).
- **Consistent aggregation**: metrics are aggregated based on the selected dimensions (see [How Sidemantic Works](how-it-works)).
- **Multiple inputs**: author models in YAML, SQL DDL, or Python.

## What It Is Not

- Not an ETL tool or transformation framework (it runs *on top of* your warehouse).
- Not a BI tool (it powers BI tools and programmatic queries).
- Not a query builder that lets you write arbitrary SQL aggregations (you define metrics up front).

## Next

- [Quickstart](getting-started)
- [How Sidemantic Works](how-it-works)
- [YAML Reference](yaml-reference)