dbt + Cosmos
Orchestrate dbt with Airflow using Cosmos.
Orchestrate, run and observe your dbt workflows with Cosmos, the open-source standard for seamless dbt orchestration and model-level visibility in Apache Airflow.
The easiest way to run dbt in Airflow.
Cosmos transforms how teams run dbt by making every model, seed, and test a first-class Airflow task. Get started in minutes with just 10 lines of code and one command—no custom scripts or complex setup required.
Unified workflows
Centralized orchestration for all your data workflows.
Cosmos unifies dbt transformations with data ingestion, ML workflows, and analytics in a single Airflow orchestration layer to eliminate tool sprawl and fragmented workflows. Watch a quick overview video to see why Cosmos is the right choice.
Complete visibility
Instant model-level pipeline clarity.
Cosmos makes each dbt model an observable Airflow task with complete lineage and execution status removing debugging bottlenecks.
Take a two minute guided tour of Astro + Cosmos
See how to connect your dbt project, configure credentials, deploy with `astro dbt deploy`, and monitor model-level execution and efficient retries.
Faster recovery
Recover faster with efficient retries.
Rerunning entire pipelines wastes time and costs. Cosmos enables targeted recovery by rerunning only failed models as individual Airflow tasks.
Platform flexibility
Flexible, future-proof deployment.
Avoid vendor lock-in and infrastructure constraints. Cosmos works across Astro, MWAA, Cloud Composer, and self-hosted Airflow with dbt Core or Fusion support.
Cosmos in Action
See how the popular Jaffle Shop dbt project converts into an observable Airflow DAG with just a few lines of code in our documentation.
Organizations from every industry transform their data operations with Cosmos.
insurance
“Since adopting Astro with Cosmos, we’ve rescued countless failed dbt jobs and still met our SLAs. That wasn’t possible before. The ability to see our dbt jobs at a more atomic level using Cosmos is game changing for us.”
Josh Bickmeyer
Manager, Analytics Engineering
online dating
“Having set up dbt-core on Airflow before, I had to create a lot of internal jobs to achieve what this project does out of the box. With Cosmos and the open source community, it's been great to see best practices in place and how forward-thinking the project is. dbt artifacts automatically spitting out open-lineage datasets to feed into data-aware scheduling? Nice! It's been great and I know it will get better.”
Adam Underwood
Senior Data Engineer
Financial technology
“After testing Mage.ai, dbt Cloud, GitHub Actions, and Cosmos, our team unanimously chose Cosmos. The integration, visibility, and simplicity won us over.”
Jensen Yap
Senior Data Engineer
responsive acquisition marketing
“Cosmos significantly accelerated our Airflow adoption, dynamically creating Airflow tasks while preserving dbt model lineage effortlessly.”
Karen Connelly
Senior Manager, Business Intelligence
Master dbt + Airflow best practice
Discover expert strategies for organizing pipelines, managing dependencies, and troubleshooting workflows with Cosmos.
Watch Best Practices VideoFrequently Asked Questions
Is Cosmos open-source?
Yes. Cosmos is Apache 2.0 licensed, with an active community and regular feature releases.
How do I get started with Cosmos?
Install Cosmos, add 10 lines of code to your Airflow DAG, and run your dbt project as observable Airflow tasks. Check out getting started tutorial for step-by-step instructions
Does Cosmos support dbt Fusion?
Yes. Once you install the dbt Fusion binary, Cosmos will recognize and orchestrate Fusion projects. Dbt Fusion is ideal for large dbt projects needing faster parsing.
How do I get started with Cosmos on different platforms?
Cosmos works across all major Airflow platforms. Follow our platform-specific guides for Astro, Open Source Airflow, and Google Cloud Composer.
Do I need to manage dbt profiles?
No. Cosmos lets you run dbt projects using Airflow connections, simplifying credential and environment management.
Where can I find working examples of Cosmos DAGs?
Explore real-world implementations in the cosmos-example repository and development examples for practical DAG patterns and configurations.
