The easiest way to turn scripts into production workflows.

Dagu adds scheduling, dependencies, retries, logs, and a Web UI around your existing scripts, commands, containers, server tasks, and AI-assisted steps. No database. No message broker. No SDK rewrite.

dagu server --port 8080

Install dagu command

$curl -fsSL https://raw.githubusercontent.com/dagucloud/dagu/main/scripts/installer.sh | bash

Guided installer: adds Dagu to your PATH, sets up a background service, and creates the first admin so you can start running workflows.

Non-invasive

No SDK required. Your business logic stays untouched.

Lightweight

Single binary, no required database or broker

Command-native

Run scripts, containers, SSH tasks, and HTTP calls

Air-Gapped Ready

Runs fully offline. No external services needed.

workflow.yaml

# Existing scripts, now production workflows

name: "daily-ops"

schedule: "0 2 * * *"

steps:

- name: "backup-db"

command: "./scripts/backup-postgres.sh"

output: BACKUP_PATH

- name: "upload-backup"

command: "aws s3 cp $${BACKUP_PATH} s3://backups/"

depends: backup-db

- name: "notify"

type: http

depends: upload-backup

Common workflow step types

Shell Docker SSH S3 HTTP PostgreSQL SQLite

Design Principles

Single binary. File-backed state. Existing code stays unchanged.

Lightweight to run

Dagu is a single binary with no required database, message broker, or control-plane stack. Start on one machine and add workers only when you need them.

Durable executions for any operation

Use schedules, DAG dependencies, retries, queues, parameters, secrets, notifications, SSH steps, container steps, and distributed execution in readable YAML.

Non-invasive to business logic

Your scripts, services, SQL, containers, and operational commands stay as they are. Dagu orchestrates around them instead of forcing a framework or SDK into your codebase.

Observable by default

Every run gets status, logs, history, timing, and a visual workflow view, so jobs stop disappearing into crontabs and server log files.

Production Workflow Requirements

Operational checks teams make before adopting a workflow engine: throughput, queues, scheduling, recovery, access control, API access, and worker execution.

Thousands/day

Single-node throughput

Run thousands of workflow runs per day on one machine, depending on hardware, workflow shape, step duration, and queue settings.

Queues + workers

Scale execution safely

Use queues, concurrency limits, and distributed workers to control load and spread jobs across machines.

Queues Distributed execution

Catchup + retry

Recover scheduled work

Cron schedules, catchup, durable automatic retries, timeouts, reruns, event handler scripts, and email notifications keep failures manageable.

Scheduling Durable retries Event handlers Email notifications

Users + API

Operate as a team

Use user management, RBAC, workspaces, approvals, secrets, REST API, CLI, and webhooks for shared production workflows.

RBAC Workspaces REST API

Use Cases

A practical index of jobs that start as scripts and need production workflow controls.

Operational Use Cases

Dagu fits where operational work already exists as commands, scripts, containers, and server tasks, then needs scheduling, retries, dependencies, logs, and handoff.

One readable YAML file. Existing commands. A run history people can actually use.

Example / Hidden cron work

Cron and Legacy Script Management

Bring existing shell scripts, Python scripts, HTTP calls, and scheduled jobs into Dagu without rewriting them.

Dependencies, status, logs, retries, and history become visible in the Web UI instead of being hidden across crontabs and server log files.

The UI stays simple enough for operators, while the workflow stays concrete enough for engineers.

Daily jobs people can maintain

ETL and Data Operations

Run PostgreSQL or SQLite queries, S3 transfers, jq transforms, validation steps, and reusable sub-workflows.

Daily data workflows stay declarative, observable, and easy to retry when one step fails.

Distributed media work

Media Conversion

Run ffmpeg, thumbnail extraction, audio normalization, image processing, and other compute-heavy jobs across workers.

Conversion work can run across distributed workers while status, history, logs, and artifacts stay in one persistence layer for monitoring, debugging, and retries.

Scheduled remote jobs

Infrastructure and Server Automation

Coordinate SSH backups, cleanup jobs, deploy scripts, patch windows, precondition checks, and lifecycle hooks.

Remote operations get schedules, retries, notifications, and per-step logs without requiring operators to SSH into servers for every recovery.

Container-native pipelines

Container and Kubernetes Workflows

Compose workflows where each step can run a Docker image, Kubernetes Job, shell command, or validation step.

Image-based tasks can be routed to the right workers without building a custom control plane around containers.

Non-engineer operations

Customer Support Automation

Run diagnostics, account repair jobs, data checks, and approval-gated support actions from a simple Web UI.

Non-engineers can operate reviewed workflows while engineers keep commands, logs, and results traceable.

Small devices, visible runs

IoT and Edge Workflows

Run sensor polling, local cleanup, offline sync, health checks, and device maintenance jobs on small devices.

The single binary and file-backed state work well on edge devices while still providing visibility through the Web UI.

YAML agents can read and change

AI Agent Automation

Use AI agents to write, update, debug, and repair workflows because the operational contract is plain YAML.

Agent-generated changes stay reviewable and observable in the same workflow system humans already operate.

Common thread

Plain YAMLAny commandDocker and Kubernetes JobsSSHSchedulesRetriesLogsNotifications

Common Workflow Patterns

Bring scripts, scheduled jobs, server tasks, and controlled automation into one workflow engine.

Health Check

SSH Backup

Notify

Script Workflows

Turn existing shell scripts, Docker commands, SSH tasks, and HTTP calls into reliable workflows.

1Keep existing scripts and commands intact
2Run containers, SSH tasks, and HTTP steps in one DAG
3Use dependencies instead of fragile command chains
4Retry failed steps with clear logs and history

workflow.yaml

steps:
  - name: health-check
    command: curl -sf http://app:8080/health

  - name: backup
    type: ssh
    config:
      host: db-server
      user: admin
    command: pg_dump mydb > /backups/daily.sql

  - name: notify
    type: http
    config:
      url: "https://hooks.slack.com/..."
      method: POST
    body: '{"text": "Backup complete"}'

Workflow Operator for Slack & Telegram

Persistent AI operator for Slack and Telegram.
Debug failures, approve actions, and recover incidents without leaving the conversation.

Dagu

Message...

Workflow engine features for real operations

Dagu focuses on the production layer around your existing work: schedules, dependencies, retries, logs, queues, and controlled execution.

Command Steps

Run existing shell scripts and commands without rewriting business logic.

Learn more

Nested Workflows

Compose larger operations from smaller reusable workflows.

Learn more

Conditional Steps

Branch workflow execution based on status, outputs, and runtime conditions.

Learn more

Automatic Retries

Retry failed steps with configurable policies and backoff.

Learn more

Durable Execution

Re-run failed workflows predictably while preserving run history and logs.

Learn more

Advanced Scheduler

Run recurring workflows with cron syntax, timezone support, and manual controls.

Learn more

Queue Management

Control concurrency and prevent operational jobs from overwhelming shared systems.

Learn more

Web UI

Inspect DAGs, runs, logs, history, and workflow definitions from one place.

Learn more

Container Steps

Run workflow steps in Docker containers when isolation or packaged tooling matters.

Learn more

Remote Execution

Run commands on remote servers through SSH while keeping status and logs in Dagu.

Learn more

Webhook Management

Trigger workflows from external systems without adding another service layer.

Learn more

Approval Gates

Pause sensitive workflow steps for human approval before they continue.

Learn more

Email Notifications

Notify people when important workflow runs succeed, fail, or require attention.

Learn more

Secret Management

Use secrets in workflow steps while masking sensitive values in logs.

Learn more

OpenTelemetry

Export traces so workflow runs can fit into your production observability stack.

Learn more

Prometheus Metrics

Expose workflow metrics for monitoring, dashboards, and alerts.

Learn more

Distributed Execution

Add workers when jobs need to run across multiple machines.

Learn more

Lifecycle Handlers

Run follow-up logic on workflow success, failure, or exit.

Learn more

Read workflow docs

Quickstart Guide

Install Dagu with the guided wizard, then continue in the full installation guide or quickstart docs.

Install dagu command

The script installers are the recommended path. Homebrew, npm, and Docker remain available for binary-only or container installs.

Mac/Linux Terminal

$curl -fsSL https://raw.githubusercontent.com/dagucloud/dagu/main/scripts/installer.sh | bash

✓ Guided installer ready

Next steps

The guided installer can finish the first-run setup for you.

# What the installer can do

Add Dagu to your PATH

Set up a background service

Create and verify the first admin

Installation Guide Read Quickstart

Project Community

Discuss usage, report issues, and follow development.

GitHub

Star the repo and follow the project.

Discord

Chat with the community