Getting Started with SemanticFed
Welcome to SemanticFed — the data federation platform that lets you query all your data in place, behind one governed semantic layer. This guide orients you around the core workflow.
SemanticFed is launching in 2026 and is currently in waitlist mode. Join the waitlist at semanticfed.com. Capabilities marked roadmap are committed but not yet generally available.
The core workflow
1. Connect a source 🔌
Register a connector for the kind of system you're adding (PostgreSQL, MySQL, MongoDB, Snowflake, BigQuery, Redshift, Databricks, ClickHouse, S3/Parquet, or a REST/GraphQL API), then register a data source — an endpoint, a credential reference, a push-down toggle, and tags. A connection probe reports health and latency and moves the source to CONNECTED.
2. Model your business 🧱
Create a semantic model and add entities bound to your sources. Define dimensions to slice by and metrics as aggregations, so "active revenue" is a documented, typed object resolved the same way everywhere. Add relationships so the engine can join across entities — and across sources.
3. Query across every source 🔎
Write a saved query that joins your operational database, warehouse facts, and lake data in one statement. The cost-aware planner pushes predicates and aggregations down to each source and stitches a single result. Each query run records a plan summary, sources touched, duration, row count, and cache status. (The live federated engine is on the roadmap; the model, query, run, and plan records exist today.)
4. Govern access 🛡️
Define an access policy once — a row filter, a column mask, or entity-level access — attached to a model or entity. It is enforced on every federated query, whatever source the data lives in. Every run is attributable in the audit trail, and lineage traces a dataset back through its views to the physical sources.
5. Ground your AI 🤖
Point an agent or copilot at the semantic layer via the MCP server or an SDK. The agent reasons over documented metrics and dimensions, with the same access policy as a human, and every answer resolves to an inspectable plan.
Continue to the Quick Start to install the CLI and run your first command, read the Architecture overview to understand how SemanticFed is put together, or see Environment Selection for BURDENOFF_ENV usage.