Dagu
Manage complex jobs with a powerful, local-first workflow engine. Define pipelines in simple YAML, run any script, and monitor everything via a built-in web UI.
dagu.sh is a powerful, local-first workflow engine for managing complex jobs without the typical operational overhead. It's designed as a zero-ops alternative to complex orchestrators, requiring no databases or external dependencies. Everything you need is packed into a single, self-contained binary with a built-in web interface for monitoring and management. You define your entire pipeline using simple YAML files.
This approach provides several key advantages:
- Zero Intrusion: Your code and scripts remain untouched. There's no need to install or integrate any SDKs into your applications.
- Completely Portable: As a single binary, it can run anywhere—on your laptop, a server, or in a completely air-gapped environment.
- Language Agnostic: Execute any script or program, regardless of the programming language it's written in.
- Integrated Web UI: Comes with a modern interface for monitoring, managing, and observing your workflows right out of the box.
Beyond standard data pipelines and CI/CD automation, it excels at running autonomous AI-agent workflows. You can build pipelines that allow AI agents to execute commands, read files, and reason about problems. It includes features like human-in-the-loop approval for sensitive operations, security policies to control agent actions, and support for models from OpenAI, Google, or local instances.
Directory Structure
docker-compose.yml
services:
dagu:
image: ghcr.io/dagu-org/dagu:latest
container_name: dagu
ports:
- "8080:8080"
environment:
- DAGU_HOST=0.0.0.0
- DAGU_PORT=8080
- DAGU_IS_BASICAUTH=1
- DAGU_ADMIN_ID=${DAGU_ADMIN_ID}
- DAGU_ADMIN_PASSWORD=${DAGU_ADMIN_PASSWORD}
volumes:
- ./dags:/home/dagu/dags
- ./data:/home/dagu/.dagu
restart: unless-stopped.env
DAGU_ADMIN_ID=admin
DAGU_ADMIN_PASSWORD=your_super_secret_passwordTags:
Similar to Dagu