SLayer

SLayer is a semantic layer for AI agents and people that query data warehouses. It sits between the agent and the database and handles read-only connectivity, SQL translation, common data transformations, and row-level security, so the agent sends a structured request with governed joins instead of writing raw SQL that has to be exactly right.

Traditional semantic layers belong to the BI stack and need every metric and aggregation predefined. SLayer lets you define a column such as revenue once and query it with expressions at query time: different aggregations, ratios, and time shifts such as a year-over-year comparison, plus multi-stage queries built on earlier results. It is organized around the agent flow of search, inspect, and query, with a search tool for discovering models and a memory store for linking business context to them.

Agents, apps, and people can reach SLayer through MCP, a REST API, a CLI, a Python client, Flight SQL, or a Postgres-compatible SQL interface. It runs as a standalone service or embeds in-process as a Python library, and model definitions can be managed by hand or by an agent through its MCP tools. It supports most popular databases and introspects a datasource's schema on startup.

SLayer fits next to an existing data stack, with importers for dbt, Cube, and Ossie configurations. A bundled Jaffle Shop demo dataset lets you try it without a warehouse. It is the core of Motley's product.