Dagster
Open-source, asset-based data orchestrator with Dagster+ cloud, Compass AI analyst in Slack and Dagster+ AI that finds and fixes pipeline failures
Overview
Dagster is an open-source Python orchestrator built around data assets: you declare the tables, files and models you want, and Dagster schedules, tests and tracks their lineage and freshness. Its homepage now reads "Data your team trusts. AI that runs on it." You can self-host the Apache-2.0 project or use the hosted Dagster+. The AI features that put it on this list are Compass, a Slack-based AI data analyst that answers plain-English questions from your warehouse (publicly available since November 2025); Dagster+ AI (early-access preview), which summarizes run failures, finds root causes, scans deployments for failure patterns and can send a coding agent to open a fix as a GitHub pull request; and a dg CLI and MCP server (dagster-mcp) that give coding agents validated actions. In July 2026 Prefect agreed to acquire Dagster Labs; Dagster's FAQ says the deal was expected to close in the coming weeks and that pricing, contracts and the open-source project stay unchanged. Also from Prefect: Prefect and FastMCP.
The Verdict
Who Should Use Dagster?
Best For
- Data teams that think in assets, lineage and freshness
- Python teams that want testable, typed pipelines
- Platforms that combine dbt, ML and AI pipelines
Not Ideal For
- Teams that only need a simple cron-style scheduler
- Buyers wary of a pending change of ownership
What's Great
- Reviewers report much less manual pipeline work
- Faster delivery and more trust in pipelines
- Clear abstractions; seen as more modern than Airflow
Watch Out For
- Steep learning curve
- Retries restart from the first step; slow UI reported
- Few reviews so far on G2
Pricing
View all features & details
Key Features
- Software-defined assets with lineage, partitions and freshness checks
- Dagster+ hosted (serverless or hybrid) with catalog search
- Compass AI data analyst in Slack
- Dagster+ AI: failure chat, proactive monitoring, auto-fix pull requests (preview)
- dg CLI and dagster-mcp for coding agents
Company
- Created by Nick Schrock; Dagster Labs (formerly Elementl), CEO Pete Hunt
- Prefect agreed to acquire Dagster Labs (announced Jul 13, 2026)
- About 40 Dagster staff to join Prefect
Real-World Usage
Customer Stories
- smava: zero downtime with Dagster
- US Foods: 99.996% uptime
- Clippd: 8+ hours a week saved
How It Compares
| Feature | Dagster | Prefect |
|---|---|---|
| Model | Data assets and lineage | Python flows and tasks |
| AI feature | Compass, Dagster+ AI | Horizon, FastMCP |
| Cloud from | $10/mo (Solo) | Free (Hobby) |
| License | Apache-2.0 | Apache-2.0 core |
Sources: Dagster · Pricing · Compass (Nov 2025) · Dagster+ AI docs · dagster-mcp · Prefect deal (Jul 2026) · Customer FAQ
Is this your tool? Add a badge