Dagu

Overview

Dagu is a self-hosted workflow engine for people who want to define jobs as YAML DAGs and manage their runs in a web UI. Its strongest case is joining existing commands and services without requiring a separate database or message broker; its free tier is generous on servers and workers but limits API keys.

It is open source under the GNU General Public License v3.0 and runs as a single binary locally or across distributed workers. That makes it a practical fit for teams comfortable operating their own workflow infrastructure, rather than buyers looking for a hosted-first service.

Dagu belongs to Workload Automation Software.

Key features

Workflows can run shell scripts, Docker containers, SSH commands, SQL queries, HTTP requests, chat completion steps, and Agent Harness steps. This range is useful when jobs already exist in different forms and need to be coordinated in one DAG, though YAML-based workflow definition will suit teams happier with configuration than a purely visual builder.

Cron scheduling includes timezone support, catchup windows, and overlap policies. Retry policies, zombie detection, lifecycle handlers, and notifications add ways to handle failures and run conditions, while workflow dependencies and event triggers cover more than calendar-driven execution. Cross-platform execution is also supported.

Built-in actions span HTTP, mail, SSH, SFTP, Git, S3, Redis, LLM, browser and computer interactions, SQL, PostgreSQL, SQLite, and DuckDB. Dagu also supports MCP connections from clients including Claude Code, Codex, Cursor, VS Code, Gemini CLI, OpenCode, Zed, and Cline. These options make it relevant to teams connecting infrastructure and AI-oriented tools in the same workflow system.

Paid Team and Pro licenses add SSO, role-based access control, audit logging, incident routing, and support to self-hosted servers. Community users can turn to documentation, GitHub Issues, and Discord, with support on those channels described as best effort. The product’s privacy policy says hosted-service data is on Google Cloud Platform and uses encryption in transit and at rest, access controls, and tenant-isolation safeguards.

Pricing

Dagu has a free Community plan and paid licenses from $50 per month. A free self-host license trial lasts 14 days and does not require a credit card.

Community is well suited to individual operators and teams that can work within two API keys and do not need the paid administrative controls. Team adds those controls across three licensed servers; Pro keeps the same listed feature set but raises the server allotment to 15. Enterprise is for organizations needing a custom server count and priority support. Support for Team and Pro is standard and handled directly by Dagu’s maintainer.

Platforms

Dagu supports API, Linux, macOS, self-hosted, web, and Windows environments. Its single-binary deployment and distributed-worker option give operators a choice between local use and broader execution, without adding a database or message broker requirement.

Who it's for

Dagu suits engineering and operations teams that want self-hosted orchestration for scripts, containers, service calls, and SQL jobs, with scheduling and run management under one roof. It is less compelling for teams that need SSO, access controls, or audit logging but want to stay on a free license, since those capabilities start with Team.

Pros and cons

• Broad execution options: workflows can combine shell, Docker, SSH, SQL, HTTP, chat completion, and Agent Harness steps.
• Lean deployment requirements: it runs as a single binary without an external database or message broker.
• Generous free infrastructure allowance: Community permits unlimited servers and workers, but is capped at two API keys.
• Paid controls are clearly separated: SSO, RBAC, audit logging, and incident routing are available on Team and Pro, but not on Community.
• Pro costs more for capacity rather than a different listed feature set: its main distinction from Team is 15 rather than 3 licensed servers.

Alternatives

Choose Kestra if you want a free platform with an Apache 2.0 release. Consider Rundeck if its free Community plan for small teams, with registration required, better matches your needs. Apache Airflow is another free, Apache-licensed option, while Luigi is a free Apache-2.0 licensed Python package without a free trial.

Windmill may fit teams looking for a free and open-source plan with stated workspace, user, and storage allowances. Prefect is worth considering if its Hobby plan’s stated limits—two users, one workspace, five deployments, and 500 minutes per month of Prefect Serverless—fit your scale. StackStorm is a free open-source choice with no paid products offered by the project. For enterprise workload automation with a free trial, consider HCL Workload Automation.

Verdict

Choose Dagu if you want open-source, self-hosted workflow orchestration that can connect existing jobs across a broad set of execution types without adding a database or message broker. Its free server and worker allowance is a strong starting point; look elsewhere or budget for Team if you need SSO, role-based access control, audit logging, or incident routing.