SLayer is an open-source, embeddable semantic layer that gives AI agents governed, shared access to warehouse data and metrics through read-only connectivity, SQL translation, and row-level security.
Adapted from @DanKornasYour AI agent needs data context—not another pile of schema dumps. SLayer is an embeddable semantic layer for AI agents and humans working with warehouse data. It helps you give agents a governed, shared surface for data and metrics by handling read-only database connectivity, SQL translation, common transformations, and row-level security. Key features: • Define once, query flexibly – reuse a column across sums, averages, ratios, time shifts, and multi-stage queries • Agent-friendly discovery – search and memory tools support the search → inspect → query flow • Multiple integration paths – interact through MCP, REST API, CLI, Python, Flight SQL, or a Postgres-based SQL API • Fits existing definitions – import configs from dbt, Cube, and Ossie • Low-friction first run – install with uv and connect Claude Code to the preloaded Jaffle Shop demo through MCP It’s open-source (MIT license). Link in the reply 👇
