How-to Guides
Task-focused walkthroughs that show FLUID working alongside the tools you already run — ingestion sources, transformation engines, orchestrators, and AI agents. Each guide is self-contained; start with the one closest to your problem.
Most of these guides use a legacy manifest shape
The YAML in these guides predates the current schema — most of it declares fluidVersion: "1.0", a version that was never published, and the build-patterns guide uses 0.3.0 — and it does not validate against the latest stable schema, 0.7.5. Each guide says so at the top. Read them for the integration patterns; for YAML that validates, use Examples, the Anatomy and the release notes.
For a step-by-step tour of the schema itself, see Examples. For field-level reference, see the Cheatsheet.
| Guide | What it covers |
|---|---|
| Source-aligned data product | Mirror an on-prem Oracle database into a governed cloud bronze layer with in-flight quality and privacy rules. |
| Source-aligned from Kafka | Continuously ingest a Kafka topic into trusted Parquet, enforcing schema, quality, and PII treatment in-stream. |
| dbt integration | Orchestrate dbt as a transformation engine and inherit its model contracts into FLUID. |
| Airflow integration | Generate contract-aware Airflow DAGs from .fluid.yml files, with no glue code. |
| Agentic access with MCP | Serve governable, auditable data to AI agents — including the 0.7.4 exposes[].mcp gateway path. |
| Data Vault 2.0 | Model hubs and satellites with dbt + FLUID, letting the orchestrator build the dependency DAG. |
| Build patterns | Tour of the canonical build-pattern styles (declarative, hybrid-reference, embedded, logical-mapping). |
| Advanced examples | The art of the possible — multi-tenant policies, feature stores, ephemeral agent products, and more. |