At a glance
| Source connectors | No built-in connectors of its own; integrations call tools such as Airbyte, Fivetran, dlt or Sling |
| Destination connectors | Through integrations and I/O managers for warehouses and storage |
| Batch or streaming | Batch and event-driven runs; not a streaming engine |
| Change data capture | Not built in; orchestrate a CDC or ingestion tool |
| Transformations | Python and SQL in assets; dbt models can be loaded as assets |
| Orchestration | Core purpose: asset-centred scheduling, automation, lineage and monitoring |
| Deployment | Self-hosted open source, or Dagster+ Serverless or Hybrid |
| Licensing | Apache License 2.0; Dagster+ by plan plus credits |
Dagster suits Python data teams that think in terms of the tables and models they produce and want lineage, data quality checks and asset-level automation at the centre of their orchestrator. The open source core is free; Dagster+ adds a managed control plane, the asset catalog, insights and alerts, billed by plan and per-materialisation credits.
Overview
Dagster is an open source platform for building, running and monitoring data assets such as tables, datasets, machine learning models and reports. You declare assets as Python functions decorated with @dg.asset, with dependencies between them, and Dagster schedules and keeps them up to date. Ops and jobs remain available for task-style work. The latest release on GitHub when checked was 1.13.29 (5 November 2026), under the Apache License 2.0.
Dagster+ is the commercial service from Dagster Labs. It runs the control plane for you, either with Serverless compute or in Hybrid mode, where an agent runs your code in your own infrastructure. Dagster+ adds features its documentation marks as Dagster+ only, including the asset catalog, insights, alerts and asset health status.
Key features
- Software-defined assets declared in Python, with an asset lineage graph
- Asset checks and data contracts for testing data as it is produced
- Declarative Automation, schedules, sensors and asset sensors
- Partitions and backfills for incremental and historical runs
- Components for configuring common integrations with less code
- Dagster+ asset catalog, insights, alerts and asset health status
- Serverless or Hybrid deployment for Dagster+
Pros and cons
Pros
- Asset-first model puts lineage and data state at the centre of the tool
- Free, Apache-licensed open source core you can self-host
- Built-in asset checks, partitions and backfills
- Published Dagster+ prices for Solo and Starter, with a 30-day trial
Cons
- The asset-based model takes time to learn for teams used to task DAGs
- Credit billing grows with the number of materialisations and op executions
- Several team features (catalog, insights, alerts) are Dagster+ only
- Fewer third-party managed hosting choices than Apache Airflow
Pricing
Dagster open source is free under the Apache License 2.0. Dagster+, the managed control plane, is billed by plan plus credits, where a credit is one asset materialisation or one op execution. As listed on dagster.io/pricing on 7 November 2026 (USD): Solo is USD 10 a month (1 user, 1 code location, 1 deployment) plus USD 0.040 per credit; Starter is USD 100 a month (up to 3 users, 5 code locations, 1 deployment) plus USD 0.035 per credit; Pro is priced by Dagster sales. Serverless compute on Solo and Starter is USD 0.010 per minute, and Hybrid deployments have no Dagster compute charge. Solo and Starter have a 30-day free trial. No region is stated; prices are listed globally.
Pricing checked on the vendor's official page on 7 November 2026. Prices change; confirm before buying.
Who it is best for
Python data teams that model pipelines as tables and models, want lineage and data checks built in, and are comfortable with either self-hosting or Dagster+.
Who should look elsewhere
- Teams standardised on Airflow or wanting a managed service from their cloud provider: see Apache Airflow vs Dagster.
- Teams that only need to load SaaS data into a warehouse: see Fivetran alternatives.
Verdict
Dagster is a strong choice when you care most about the data assets themselves: lineage, checks, partitions and automation per asset. The open source core is free and Dagster+ pricing is published for small teams, but credit billing needs watching as pipelines grow, and teams used to task-based tools should allow time for the change in model.
Frequently asked questions
Is Dagster free?
Dagster open source is free under the Apache License 2.0 and can be self-hosted. Dagster+ is paid: Solo is USD 10 a month and Starter USD 100 a month plus per-credit usage (as listed on 7 November 2026), with a 30-day trial; Pro is priced by sales.
What is a Dagster+ credit?
Dagster's pricing page counts one credit for each asset materialisation and one for each op execution, whether the op runs in a job or as part of an asset.
How is Dagster different from Prefect?
Dagster centres on assets and their lineage, while Prefect centres on Python flows and tasks. See Prefect vs Dagster.
Sources
- Dagster+ pricing
- Dagster documentation
- Dagster integrations
- Dagster repository (Apache 2.0)
- Dagster releases
Checked November 2026.
How we research reviews and what star ratings mean: our editorial method.