# Ondatra Labs

> The data runtime for teams that ship. One binary replaces ingestion, transformation, and orchestration.

One binary replaces ingestion, transformation, and orchestration.

- License: MIT
- Language: Go
- Repository: https://github.com/ondatra-labs/ondatrasql
- Discord: https://discord.gg/TQEqsa9q

## Documentation

- [DuckDB Configuration](https://ondatra.sh/duckdb-configuration/): How the runtime boots — DuckDB configuration via SQL files
- [settings.sql](https://ondatra.sh/duckdb-configuration/settings/): Execution settings — how the runtime uses DuckDB
- [secrets.sql](https://ondatra.sh/duckdb-configuration/secrets/): Access external data without adding infrastructure
- [extensions.sql](https://ondatra.sh/duckdb-configuration/extensions/): Optional capabilities for the runtime
- [catalog.sql](https://ondatra.sh/duckdb-configuration/catalog/): Where your data lives — the only required configuration
- [macros.sql](https://ondatra.sh/duckdb-configuration/macros/): Reuse SQL without templating
- [variables.sql](https://ondatra.sh/duckdb-configuration/variables/): Configure your pipeline with SQL
- [sources.sql](https://ondatra.sh/duckdb-configuration/sources/): No connectors. Just queries.
- [Getting Started](https://ondatra.sh/getting-started/): Run your first data pipeline in under 5 minutes
- [CLI Reference](https://ondatra.sh/cli/): Every command in one page
- [Concepts](https://ondatra.sh/concepts/): How OndatraSQL works — the mental model
- [Models](https://ondatra.sh/concepts/models/): Everything is a model — ingestion, transformation, and events in one system
- [Directives](https://ondatra.sh/concepts/directives/): Directives turn SQL into a data pipeline
- [Model Kinds](https://ondatra.sh/concepts/kinds/): Choose data behavior with one directive — not pipeline logic
- [Smart CDC](https://ondatra.sh/concepts/cdc/): Write once. Run incrementally forever.
- [Schema Evolution](https://ondatra.sh/concepts/schema-evolution/): No migrations. Ever.
- [Incremental Models](https://ondatra.sh/concepts/incremental/): Incremental is not a feature — it's the default
- [Dependency Graph](https://ondatra.sh/concepts/dag/): Your queries define the graph
- [Sandbox Mode](https://ondatra.sh/concepts/sandbox/): See the result before you commit it
- [Maintenance](https://ondatra.sh/concepts/maintenance/): Keep your data fast and your storage under control
- [Event Collection](https://ondatra.sh/concepts/events/): Built-in event ingestion — no Kafka, no message broker
- [Validation](https://ondatra.sh/validation/): Built-in data validation — no external tools required
- [Constraints](https://ondatra.sh/validation/constraints/): Validation is part of execution — not a separate step
- [Audits](https://ondatra.sh/validation/audits/): Detect regressions — rolled back automatically if they fail
- [Warnings](https://ondatra.sh/validation/warnings/): Non-blocking validation — observe before you enforce
- [Scripting](https://ondatra.sh/scripting/): Built-in scripting for ingestion — no Python required
- [HTTP Module](https://ondatra.sh/scripting/http/): Built-in HTTP client for API ingestion
- [Save Module](https://ondatra.sh/scripting/save/): Emit rows that become the model output
- [OAuth Module](https://ondatra.sh/scripting/oauth/): Built-in OAuth for API authentication
- [Runtime Modules](https://ondatra.sh/scripting/modules/): A built-in runtime for writing data pipelines
- [Language Reference](https://ondatra.sh/scripting/language/): A simpler alternative to Python for pipelines
- [Examples](https://ondatra.sh/examples/): Real pipelines. No infrastructure.
- [Ingest from a REST API](https://ondatra.sh/examples/ingest-rest-api/): Fetch data from a real API — no Python, no pip, no setup
- [YAML Model with Shared Library](https://ondatra.sh/examples/yaml-model/): Reusable source function with YAML configuration — one connector, many models
- [Track Product Analytics](https://ondatra.sh/examples/track-product-analytics/): Collect events from your app, transform them into metrics, and query the results
- [Build Daily Metrics](https://ondatra.sh/examples/build-daily-metrics/): Create a multi-layer SQL pipeline with incremental updates and validation
- [Incremental SQL Model](https://ondatra.sh/examples/incremental-model/): Build an incremental append model with cursor tracking
- [SCD2 Dimension](https://ondatra.sh/examples/scd2-dimension/): Track dimension history with Slowly Changing Dimension Type 2
- [Validated Model](https://ondatra.sh/examples/validated-model/): Combine constraints and audits for comprehensive data quality
- [Lineage & Metadata](https://ondatra.sh/lineage/): Built-in lineage and metadata — no external systems required

## Blueprints

- [Google Ad Manager](https://ondatra.sh/blueprints/google-ad-manager/): Fetch historical GAM reports incrementally with one YAML model and one shared source function.

## Blog

- [The Modern Data Stack Is Solving the Wrong Problem](https://ondatra.sh/blog/we-didnt-need-a-better-data-stack/): Why most teams need vertical simplicity, not horizontal scale
- [Why I Stopped Writing Data Pipelines in Python](https://ondatra.sh/blog/why-i-stopped-writing-data-pipelines-in-python/): Python made data pipelines possible. But it also made them more complex than they need to be.
- [You don't need a data warehouse — you need a catalog](https://ondatra.sh/blog/you-dont-need-a-data-warehouse/): Every data project starts with a warehouse decision. What if that decision didn't exist?

## Optional

- [About](https://ondatra.sh/about/): The story behind OndatraSQL — why I built a data runtime instead of another pipeline tool